Marine Geology 369 (2015) 100–115
Contents lists available at ScienceDirect
Marine Geology
journal homepage: www.elsevier.com/locate/margeo
Submarine landslides and fault scarps along the eastern Mediterranean
Israeli continental-slope
Oded Katz a,⁎, Einav Reuven a,b, Einat Aharonov b
a
b
Geological Survey of Israel, 30 Malkhe Israel St., Jerusalem 95501, Israel
The Institute of Earth Sciences, The Hebrew University, Edmond J. Safra Campus, Givat Ram, Jerusalem 91904, Israel
a r t i c l e
i n f o
Article history:
Received 17 February 2015
Received in revised form 23 July 2015
Accepted 1 August 2015
Available online 8 August 2015
Keywords:
Submarine landslide
Fault scarp
Salt tectonics
Eastern Mediterranean
Hazard
a b s t r a c t
The present work maps and studies the continental slope off the southeastern Mediterranean Israeli coast. Bathymetric grids with 15–50 m/pixel resolution were used to map over four hundred submarine landslides and numerous fault scarps exposed on the sea floor.
Landslide scars are found at water depth ranging between 130 m and 1000 m, where slopes exceed a critical gradient of about 4°–5°. Landslide surface areas range from 0.0024 km2 to 91 km2, where the observed size distribution has a peak (roll over) showing that the most probable landslide area is 1.6 × 10−2 km2. In general landslides
in the north of the studied area are smaller and occur at shallower depth than the southern ones. Landslides show
a hierarchical pattern, resulting from sequential, retrograding, slope-failure events and are also observed to interact with a group of faults oriented sub-parallel to the coast. These faults are a result of salt tectonic related extension, their scarps forming elongated step-like morphological features rupturing the surface of the continental
slope, as well as the deeper sea floor.
The morphology of the landslides as well as their cross cutting relation with the faults scarps, suggest that these
landslide are recent, apparently younger than 50,000 years. The triggering mechanism is not clear yet, though
several conditions which are known to promote slope instability prevail in the studied area: submarine slope gradients are close to the inferred critical slope angle; continuous sedimentation increases the load on the slope; active salt tectonic results in an overall extension and surface rupturing by normal faults; the studied area is merely
100 km away from seismogenic zones; and finally, apparent existence of gas close to the surface. Hence, it is suggested that submarine slope failure events in the studied area are also possible in the future.
© 2015 Elsevier B.V. All rights reserved.
1. Introduction
Submarine landslides are a result of submerged slope-failures, generating sediment transport from the continental shelf and upper slope
toward the deep basins (e.g., Hampton et al., 1996; Mulder and
Cochonat, 1996; McAdoo et al., 2000; Hutton and Syvitski, 2004; Lee,
2009). In general, landslides initiate as slope-material failure, followed
by down-slope movements of coherent masses of sediment on discrete
failure planes (Masson et al., 2006). Submarine landslides occur in many
environments worldwide, such as oceanic volcanoes, river deltas, submarine canyons and open slopes (Lee, 2009), and are known in the
Mediterranean Sea as well (Camerlenghi et al., 2010; Urgeles and
Camerlenghi, 2013), where they occur in various passive and active
margins settings.
Submarine landslides are recognized using bathymetric mapping and
seismic profiling according to their distinct morphological (Masson et al.,
2006; Bull et al., 2009; Chaytor et al., 2009) and structural indicators
(McAdoo et al., 2000; Twichell et al., 2009; Camerlenghi et al., 2010;
⁎ Corresponding author.
E-mail address: [email protected] (O. Katz).
http://dx.doi.org/10.1016/j.margeo.2015.08.006
0025-3227/© 2015 Elsevier B.V. All rights reserved.
Frey-Martinez, 2010; Urgeles et al., 2010; Winkelmann et al., 2010). Submarine landslides are abundant even on very shallow slopes, and those
that occur on open slopes tend to be the largest (Booth et al., 1993).
The runout of landslides is highly variable: the mass may arrest soon
after failure, it may creep, or it may turn into rapid debris or turbidity
flows (Tripsanas et al., 2008), running up to hundreds of kilometers before arresting (Shaller and Smith-Shaller, 1996; Twichell et al., 2009). For
this reason submarine landslides may endanger offshore submerged facilities (e.g., Locat and Lee, 2002; Thomas et al., 2010). In addition,
rapid underwater landslides also endanger adjacent coastal areas by generating tsunamis (e.g., Longva et al., 2003; Sultan et al., 2004; Tinti et al.,
2006; Tappin, 2010).
Given the hazard that submarine landslides pose, it is not surprising
that they are the focus of numerous studies worldwide (e.g., Masson
et al., 2006). In particular, the densely populated Mediterranean coast
has been subject to intense study. Recently, Urgeles and Camerlenghi
(2013 and reference therein) compiled a catalog of close to 700 late
Quaternary submarine landslides around the Mediterranean, investigating landslide size distributions and triggering mechanisms. Urgeles
and Camerlenghi (2013), found that landslides currently exposed on
the Mediterranean sea floor were not necessarily triggered by
O. Katz et al. / Marine Geology 369 (2015) 100–115
earthquakes, rather fluids play a major role in submarine slope instability (observed slope failure is often related to pockmarks and mud volcanoes). They also suggested that climate-induced stress changes
(sedimentary load, sea level, bottom temperature effect on fluid flow,
gas hydrate, and gas systems) during the last deglaciation phase, may
have had a significant effect on submarine slope stability. However,
the landslide catalog presented in the above works includes only the
large size-range of landslides and is incomplete for landslides smaller
than 1 km3 (Urgeles and Camerlenghi, 2013). In addition, the landslides
relation to submarine faults was not studied, perhaps because faults do
not play as important role for large landslides as they do for smaller ones
(this relationship will be addressed in the present work).
The present work focuses on medium to small underwater landslides, a size range that is rarely mapped underwater and was not
mapped in the Mediterranean previously (Urgeles and Camerlenghi,
2013). We center our study on the continental slope off the southeastern Mediterranean Israeli coast (Fig. 1), and use newly released highresolution bathymetric grids to map in detail landslides and fault. The
resulting landslides catalog includes, for the first time around the Mediterranean, landslide size-range smaller than 1 km3, and records the
landslides spatial relation to submarine faults. The objectives of the
work are: to map and analyze the landslide sizes and spatial distributions in the studied area, to study the mechanisms of submarine slope
failure considering also its possible relation to the faults, to assess the
time frame for the landsliding activity and then, to use these accomplishments to get a better understanding of the submarine landslide
hazard relevant to the densely populated eastern Mediterranean shore.
The high-resolution mapping of submarine landslides and faults
gives a unique opportunity to study how deep-seated salt tectonic affect
temporally and spatially surface processes shaping passive margins.
Another aim of global implications is mapping the population of
small-to-medium sizes landslide in the Levant basin. This size range of
landslides is not studied routinely, and we aim to characterize their
unique locations, triggers, ages, activity, and distributions.
1.1. Regional setting
The studied coast (Fig. 1) is controlled by northward along-shore
transport of sediments from the Nile Delta (Stanley et al., 1998;
Brenner, 2003). During Pliocene–Quaternary a 1–2 km thick wedge of
mainly Nile-derived clay-rich siliciclastic sediments accumulated in
the eastern Mediterranean over sandy turbidities and Messinian evaporites (e.g., Gvirtzman and Buchbinder, 1978; Tibor et al., 1992; Ben-Gai
et al., 2005). Recent sedimentation rates recorded on the studied continental slope (up to water-depth of 900 m), based on cores, are 25 cm/ky
to 130 cm/ky (Luz and Perelis-Grossowicz, 1980; Schilman et al., 2001;
Hamann et al., 2008; Kuhnt et al., 2008; Almogi-Labin et al., 2009). Sedimentation rates recorded for the deeper part of the SE Mediterranean
101
(beyond the continental slope) are significantly lower, b 5 cm/ky
(Calvert and Fontugne, 2001).
The current submarine offshore morphology in the study area consists of up to 200 m deep shelf that is up to 20 km wide in the south
and narrows northwards. Slope gradients on the shelf are generally
small, not more than 2° (see Fig. 2 for water depths and slope gradients).
Further to the west, on the continental slope, between 200 m and
1000 m water depth, the slope steepens to 4° or steeper values in
some locations. The continental slope consists of two lateral provinces
(Fig. 1): The submarine canyons province in its northern part (north
of Latitude 32.75°N) (not studied here), and the open slope province,
hereafter the studied area, further south of this Latitude.
Two main types of mass movements are observed along the open
slope: The first type consists of very large slumps, over 1 km thick,
with their primary sliding-plane located within the Messinian evaporites (Almagor and Garfunkel, 1979; Almagor, 1984; Cartwright and
Jackson, 2008). The internal deformation within these landslides, also
known as ‘disturbances’ (Garfunkel, 1984), consists of landward block
rotation on listric growth faults rooted at the underlying evaporites.
Two sites across the studied continental slope are believed to reveal
this kind of very large gravitational collapse (Figs. 1, 2): the Palmahim
and Dor Disturbances (Mart et al., 1978; Almagor and Garfunkel,
1979; Garfunkel et al., 1979; Roter, 2011). Other similar disturbances
are known to the south, namely the Bardawil and Gaza (Garfunkel,
1984).
The second type of submarine mass movements along the open
slope consists of small to medium size landslides (b10−2–101 km2),
which were identified previously off coast Israel (Almagor and
Garfunkel, 1979; Frey-Martinez et al., 2005), but were not comprehensibly mapped or studied previously, there or elsewhere in the Mediterranean (Urgeles and Camerlenghi, 2013). These small to medium
submarine landslide are the subject of the current work. In addition,
north–south elongated morphological step structures are also seen to
outcrop on the sea floor in the studied area. The structures are recognized as surficial scarps of growth faults that are a result of subsurface
salt tectonic related to the buried Messinian evaporitic sequence
(Garfunkel, 1984; Baudon and Cartwright, 2008; Gvirtzman et al.,
2015). This subsurface salt tectonic results in extensional deformation
of the overlaying Pliocene–Quaternary sequence near the continental
slope (Ben-Avraham, 1978; Gradmann et al., 2005; Mart and Ryan,
2007; Cartwright and Jackson, 2008; Cartwright et al., 2012).
1.2. This work
Our work consists of several steps: First we map submarine landslides on the Israeli continental slope from newly available bathymetric
maps, focusing on the small to medium landslides. The large number of
identified landslides allows us to perform analysis of their statistical
Fig. 1. Perspective view of the studied area (studied area is marked by a dashes ellipsoid). Gray coloring is the slope angle based on 50 m bathymetric grid (legend is shown below the
figure). Location is shown in the inset, where the studied area is marked by a white rectangular.
102
O. Katz et al. / Marine Geology 369 (2015) 100–115
Fig. 2. Landslides and fault scarps mapped in the studied area (for Primary, Secondary and Tertiary landslides, see text). (a) shown on top of the bathymetric grid and (b) shown on top of
the slope-angle grid (in this map landslides are not divided according to their hierarchy). A map of the landslides and fault scarps without a background and a map of the landslides and
fault scarps shown on top of the slope-angle shaded grid are part of the Supplementary data (Fig. S1).
distribution and their spatial variability. The size distribution of the submarine landslides mapped and analyzed in this work was overlooked in
many previous works targeting the studied area and elsewhere (e.g.,
Lee, 2009 and reference there in; Urgeles and Camerlenghi, 2013).
After studying the landslides, in order to better understand the landslide
system we next map the submarine fault scarps that outcrop on the sea
floor and analyze cross-cutting relations between faulting and landsliding and among the different landsliding events. Our third step consists
of augmenting our bathymetric analysis with seismic cross sections,
aiming to identify the subsurface structures controlling the morphology
and locations of the landslides and faults, to pinpoint the mechanical relations between landslides and faults.
After describing all this newly obtained data, we suggest possible
triggering mechanisms for the landslides and estimate their current activity, setting the stage for a more quantitative analysis of the triggering
mechanisms and the critical conditions for landsliding along the Israeli
continental slope, as well as enabling hazard analysis related to submarine slope failure.
This work thus contributes both new data and new insights: data revealing landslide sizes and locations in an important region of the Levant basin, data revealing fault scarps in this region and their relations
to landslides and insights relating to the mechanical controls on
observed landslide morphologies and distributions, which may be
applicable to other unstable slopes around the world. The landslides
mapped in the frame of this work are smaller than previously mapped
in the Mediterranean (Urgeles and Camerlenghi, 2013) and thus are
complementary to the data sets and the ideas that currently exist.
2. Methods
For mapping the submarine landscape we used bathymetric grids
with pixel resolution of 15 m (pixel area is 225 m2) till water depth of
about 700 m, acquired during 2000–2006 (Sade, 2007; Sade et al.,
2007) and of 50 m (pixel area is 2500 m2) till water depth of over
1700 m, acquired in 2010 (Tibor et al., 2013). Landslides were identified
in the current work, using their distinct head-scar morphology, which is
generally arc-like in map view and presents steeper slope (N 10°) than
the undisturbed submarine slope (b5°). The extent of the landslides
down slope to their distal part was recognized using their side-scar
morphology and by identifying their deposit. Thus we mapped only
landslides which are currently morphologically expressed on the
surface of the sea floor. Landslides were mapped manually using Global
O. Katz et al. / Marine Geology 369 (2015) 100–115
103
Fig. 2 (continued).
Mapper v13.2 as polygons. A single polygon separately encloses each individual recognized landslide. Relevant spatial and bathymetric characteristics (see below) of each individual mapped landslide were
measured directly using the grids and the mapped polygons, and documented in Table S1 (in supplementary data). These characteristics include: Location of the landslide, marked by the (global) geographical
coordination of the head scar center and the Section: north or south of
latitude 32.43°N; Area of the landslide, calculated as the area enclosed
by the mapped polygon that includes the scar and the deposit (similar
to Korup, 2005; ten Brink et al., 2006); Hierarchy of the landslide, ‘Primary’, ‘Secondary’, which is a landslide that develops from the scar of a Primary one, and ‘Tertiary’, which is a landslide that develops from the scar
of a Secondary one; Minimal and maximal water depth of the landslide,
measured at the head scar and the toe of each landslide, respectively;
Scar height of each landslide, calculated as the elevation difference between the top and base of the scar (scar is defined as areas where the
slope is N 10°); Type of landslide: slump (where deposited material is
present at the base of the landslides, McAdoo et al., 2000; Hungr et al.,
2014) vs. disrupted debris avalanche (where the deposited material liquefied and flowed down the slope); The field relation of each landslide
with the fault scarps, whether a landslide is triggered from the fault
scarp, covers the fault scarp or is cut by it; The slope gradient of the
landslides: For analyzing the slope-gradient of the landslides' scar
heads (steepest and upper most part of the scars), of their deposits
(slope material that has been displaced and deposited below the scar)
and of their toes (lowermost part of the deposit), these areas within
56 selected landslides where enclosed by polygons. Each polygon
hosts numerous pixels and thus contains many data points representing
the local gradients. The mode (most frequent value) of the data points
was then calculated to present the slope value.
Faults were identified according to the steep topography
of their scarps, N10°, and their overall linearity, in most cases
south - north. The scarps mark the sites of surface rupturing (see
below). Faults were manually mapped as polylines using Global
Mapper v13.2.
In addition to bathymetric mapping we investigated several 2D seismic reflection sections acquired for oil and gas exploration (on 2001 by
TGS-NOPEC Geophysical Company L.P). The cross sections were chosen
to reveal the landslides' base and the structural control. In addition,
some seismic sections were taken through surface fault scarps, in
order to see the extent, structure, and root of these faults. Seismic interpretation was carried out in the frame of this work at the seismic interpretation lab of the Geological Survey of Israel, using Kingdom software
(by HIS).
104
O. Katz et al. / Marine Geology 369 (2015) 100–115
Table 1
Slope angle of the scar-head, deposit and toe of selected landslides. Landslides are grouped
according to their location (North vs. South) and hierarchy (primary, P, secondary, S or tertiary, T). n is the number of landslides used for the calculation. Range, average and St. Dev.
are the range, average and standard deviation of the slope mode (most frequent values)
calculated over n landslides (the practice of mode evaluation in each landslide is detailed
in the Methods section). Details of each selected landslide used in Table 1 appear in
Table S1 (Supplementary data).
Description
Scar Head
Deposit
Toe
All
North
South
P
S
T
All
North
South
All
n
Range
(°)
Average
(°)
St. Dev.
(°)
43
25
18
6
26
11
10
5
5
3
5–26
5–17
9–26
5–26
7–21
7–13
2–9
2–6
3–9
1–2
13
11
16
14
13
12
5
4
6
1
4
3
4
7
4
3
2
2
2
1
3. Results
3.1. Landslides morphology
We mapped 447 landslides detectible on the sea floor of the studied
continental slope (Fig. 2). The characteristics of each individual mapped
landslide were documented (Table S1—Supplementary data) to enable
further spatial and statistical analysis of this landslide system as well
as comparison with other submarine landslide inventories worldwide.
Failed slopes cover more than 440 km2, which comprises more than
20% of the studied continental slope, from water depth of about 130 m
(shallowest scar) to more than 1100 m (lowest mapable deposit). The
height of scars ranges from less than 10 m to 80 m, while their width
(in the slope normal direction) ranges from tens of meters to a few kilometers. The slope gradient mode in the scars ranges between 11°–16°
(Table 1), it may locally exceed 25° within the scar. The distinct circular
scar morphology common to many landslides of the study area suggests
that these landslides are of the slump type (Hungr et al., 2014). Following slope failure, the failed material either kept some of its consistency
and deposited down slope relatively close to the scar at a distance of
up to 1 km (e.g., Fig. 3) forming a typical toe structure, or disintegrated,
probably liquefied and moved down slope as disrupted debris avalanche away from the scar, to a distance of 10 km or more (e.g.,
Fig. 4). Note that sliding distance is also affected by sliding volume
(e.g., Issler et al., 2005; Goren and Aharonov, 2007, 2008 and references
therein). It is more common to find debris avalanches in the south of the
studied area, and slumps in the north. The difference in the landslide
nature between north and south will be discussed below. Some landslides show characteristic of slides rather than slumps (Fig. 5), i.e., sliding on roughly planner surface (Hungr et al., 2014). These are in general
situated on the upper part of the continental slope close to the shelf
break and typically have primary scars that are up to 50 m high and
up to a few kilometers wide (whereas slumps tend to have much
narrower scars). The large width of the primary scars (i.e., scar front)
Fig. 3. A Simple Landslide developed from a fault scarp (For location see Fig. 2a). The slope material moved from the scar to the toe of the landslide. (a) Map view showing bathymetry;
(b) Map view, same extent as in a, showing slope angle; (c) Perspective view, showing slope angle; (d) Topographic profile through the landslide (a–a′) in comparison to the topography of
the intact slope south of the landslide (b–b′). The location of the profiles are marked in b. (e) Inset of the landslide scar topography in comparison to intact slope showing sites of missing
and excess of material.
O. Katz et al. / Marine Geology 369 (2015) 100–115
might be a result of laterally coalesced multiple-failures that have
moved down on a sliding plane located at similar depth. Failed material
forms rough surface area down slope from the scar. Secondary failure
occurs along the primary scar front (Fig. 5).
We used seismic lines to study the general deep structure of the
landslides. Mostly the available lines did not have sufficient resolution
to study the landslides in detail. Since our seismic resolution is estimated to about 30 m, the very fact that we cannot resolve the landslides on
seismic lines reveals that most studied, exposed on the surface, landslides are shallower than ca. 50 m. A cross section along a landslide
from the southern part of the studied area presents a deposit thickness
of up to 100 m at the landslide toe located about 12 km down slope from
the head scar which is clearly seen on the seismic line (Fig. 6, location of
the section given in Fig. 2a). The surface between the head scar and the
toe is rough and probably served as the source for the landslide material,
where the surface at the toe is smooth as result of the plastic nature of
the material deposited in the toe.
Landslides follow either a ‘simple’ (Fig. 3) or a ‘complex’ (Fig. 7)
nature. Simple landslides are defined as those that show evidence of a
single slope-failure event, continuous, almost undisturbed, arc-like
scar shape and a well constrained deposit. Complex landslides, on the
other hand, are defined as those that show a hierarchical pattern,
resulting from sequential slope-failure events. These landslides comprise a primary landslide, developed within the intact slope, and a few
secondary landslides, developed in the over steepened (N10°) head
scars of the primary one. In general, secondary landsliding occurs in
the upslope direction forming a retrograding landsliding system (e.g.,
Katz et al., 2014), where secondary landslides are smaller than the
parent ones (Figs. 7d, 8a). In places secondary landsliding is also lateral
(develops in the over steepened side scars). Tertiary landsliding also
takes place in the scars of the secondary landslides (Fig. 7). Primary
(complex and simple), secondary and tertiary landslides comprise
89%, 10% and 1% of the mapped landslides area, respectively (Table S1).
A distinct change in landslide morphologies occurs around Latitude
32.43°N (Figs. 1, 2). Landslides south of this latitude are larger
(Fig. 8a) and show a more plastic nature with lobe-like deposit material
extending down slope to distances greater than 10 km from the head
scar (Figs. 4, 6). North of this latitude landslides are smaller and less
disrupted, i.e., deposit material is located closer to the head scar. In addition landslides in the south initiate in deeper water, the top of the
105
scars at the southern part of the open slope are found at a water depth
of 350 m–400 m (Fig. 8c) where local slope exceeds 4° (Fig. 2b), while
in the northern part, around Dor disturbance, the shallowest scars are
found in the vicinity of the shelf-break (100–200 m depth), and closer
to the shore (Fig. 2). In the north the steepest part of the slopes, which
has a major control on landslide locations, is at shallower water depth
(Fig. 2b). Landslide density, measured as the number of landslides per
1′ (~1.85 km) wide E–W slope-stretch, increases northwards, where it
is distinctively higher north of Latitude 32.43°N (Fig. 9).
We next measured slope angles at various locations along scars and
deposits, and note the mode of slope angle: We measure the head scars
of 43 landslides (primary, secondary and tertiary) across the studied
area (using the bathymetric grid as explained in the Methods section;
Table S1): 25 landslides from the northern area show an average head
scarp slope mode of 11° ± 3°. 18 landslides from the southern area
show an average head scarp slope mode of 16° ± 4° (Table 1). From 6
primary, 26 secondary and 11 tertiary landslides representing south
and north combined we found that the average head scarp slope
mode does not depend on landslide generation and is 12°–14° for all
three generations. Next, we measured the slope angle mode of the
deposited material, close to the scar, of 5 landslides from the northern
and 5 from the southern areas, and found that the average modes of
the deposit slopes were 4° ± 2° and 6° ± 2°, respectively. Finally, we
also measured the slope angle of the deposited material at the toe of 3
landslides, which were selected since they have typical undisrupted
lobe-like deposits (e.g., Fig. 6). The slope angle mode at those toes is
1° ± 1°, which is much lower than the slope of the deposit material
found near the scar, which is reported above to be about 5°.
3.2. Landslides size and spatial distributions
Sizes of landslides, measured as their surface area (A), ranged from
0.0024 to 91 km2 (mode is 0.01 km2 and median is 0.09 km2). The
area of the smallest mapped landslide is an order of magnitude larger
than the area of the pixels of the used bathymetry map (till water
depth of 700 m) and thus we consider our inventory as practically complete. N–S difference in the nature of the landslides is reflected also in
the area range: In the south area ranges between 0.0077 and 91 km2
(n = 113), where in the north landslide area ranges between 0.0024
and 9 km2 (n = 334). Fig. 8a demonstrates the general reduction in
Fig. 4. Slope angle grid showing a series of landslides with liquefied toes developed at the southern end of the studied area (for location see Fig. 2a). Note the local steepening (darkening of
the gray colors) takes place where the scar heads are, at a depth of close to 400 m.
106
O. Katz et al. / Marine Geology 369 (2015) 100–115
Fig. 5. Landslide scars, marked by Sl, and fault scarps, marked by FS, exposed on the continental slope (for location see Fig. 2a). (a) Map view showing slope angle (depth is marked is 100 m
intervals); (b) Perspective view showing slope angle. Note that unlike the fault scarps, the landslide scars host secondary landslides. Shown in frames are landslides that appear in detail in
Fig. 13. Dotted surface at the western side of the images are artifacts.
landslide area going from south to north. This trend of northwards decreasing landslide sizes is strongest for the primary landslides
(Fig. 8b), because these are the ones been affected by the vary regional
topography where the secondary landslides are affected by the local
slope of the primary landslide scar.
The thickness of the landslides (t) can be roughly estimated only in a
few sites using the bathymetry (Fig. 3a) and seismic cross-sections
(Fig. 6), and it is found to be on the order of 10 m to 100 m.
Measured individual landslide surface areas (Table S1) were plotted
on a non-cumulative probability density (p) vs. magnitude (individual
landslide surface area, AL) diagram, following Malamud et al. (2004):
In order to compare with the Mediterranean submarine landslide
population studied by Urgeles and Camerlenghi (2013), the cumulative
area distribution of the 447 mapped landslides was also plotted
(Fig. 10b). The observed cumulative size distribution flattens out for
small landslide sizes, while the large landslides sizes can be well fit by
a power law relationship, so that p ~ Aθ, where θ = −0.7. This exponent
is similar to that found by Urgeles and Camerlenghi (2013), θ = −0.8.
However, their exponent fits a sub-set of their landslide population,
those with sizes of 108–1011 m2, which are beyond the sizes of the landslide population studied in the current work (103–108 m2).
3.3. Faults scarps
1 δNL
p¼
N δAL
ð1Þ
where N is the total number of landslides in the inventory and δNL is the
number of landslides with area between AL and AL + δAL. Fig. 10a shows
the probability p to find a landslide of size A is best fit by a double Pareto
distribution or an inverse Gamma distribution, as explained and shown
for other landslide size distributions in Malamud et al. (2004). The
observed size distribution has a peak (roll over), showing that the
most probable slide area is 1.6 × 104 m2. For landslides larger than
this most probable landslide size, the distribution can be well fit by a
power law relationship, so that p ~ Aα, where α = − 1.6 over the
large landslide part of the population (Fig. 10a).
Faults in the studied area have been mapped and discussed in previous works (Garfunkel, 1984; Baudon and Cartwright, 2008; Gvirtzman
et al., 2015), using both bathymetric and seismic data, mainly analyzing
their role in creating the rotated-blocks structure within the Palmahim,
Dor and other disturbances along the eastern Mediterranean (Almagor,
1984). Generally the faults were described (Garfunkel, 1984) as coast
parallel, mostly dipping basin-wards and listric, i.e., they dip steeply
close to the sea floor and flatten downwards. Seismically mapped strata
reveals that the faults are post Messinian and contemporaneous with
sedimentation. Many of them rupture the sea floor, forming a steplike morphological expression which suggests that they are still active
(Garfunkel, 1984). Here we present a more comprehensive surface
O. Katz et al. / Marine Geology 369 (2015) 100–115
107
Fig. 6. (a) A seismic line (Line 2011-IS by TGS-NOPEC Geophysical Company L.P.) crosses a landslide at the southern part of the studied area (for location see Fig. 2a) and demonstrates the
thickness of the landslide deposit at the surface. The line also reveals the absence of faults at the subsurface of the southern part of the studied area; (b) Inset of the landslide (marked by
dashed rectangular on a).
mapping of the scarps of the submarine growth faults, done using the
high-resolution bathymetry while paying special attention to morphological relations with landslides.
We mapped in the studied area numerous individual fault-scarp segments (Fig. 2). The scarps form elongated step-like morphological features rupturing the surface of the continental slope (Figs. 3, 5) and
abyssal plane. Segment lengths range from 1 to more than 10 km
long, with scarps that are up to 70 m high and slopes within the scarps
that are up to 20°.
A close correlation exists between scarps mapped using the bathymetry and faults revealed in the seismic lines (Fig. 11). Fault scarps are
exposed north of the Palmahim disturbance only. Comparison of seismic lines from north and south of Palmahim disturbance reveals abundant faulting north of Palmahim (Fig. 11) and lack of any faulting south
of it (Fig. 6).
Two groups of faults can be distinguished. The first group consists of
normal faults that are sub-parallel to the continental slope with a downdisplaced (hanging) western block. Up to 10 sub-parallel fault segments, rupturing the continental slope, can be traced along a single
slope-normal line cutting from water depth of 200 m to ca 1000 m
(Fig. 2). The lateral offset between neighboring faults ranges from less
than one to a few kilometers. The second group consists of faults that
are semi radial about the Dor disturbance (Fig. 2). Some of these faults
scarps face each other and form elongated Graben like structures
(Fig. 12). The maximum distance between adjacent faults of the second
group is up to two kilometers.
3.4. Field relation between the landslides and the growth faults
Landslides and fault scarps overlap spatially and interact at the
northern part of the open slope province only (Fig. 2, north of
32.43°N). Further south, up to 32.23°N, faults rupture the surface only
west of the continental slope and thus do not interact with the landslides that initiate on the slope above. Further south, past of the
Palmahim disturbance, faults are not exposed (Fig. 2) or observed in
the subsurface above the Messinian (Fig. 6).
84 out of the 447 mapped landslides reveal cross-cutting relations
with faults scarp. Where they occur, the cross-cutting relations between
landslides and faults scarps are variable and testify to a multi-temporal
geological history where landsliding and surface rupturing by active
faulting post and predate one another. We can identify three variants
(Table S1): (a) 55 landslides were found to initiate from fault scarps,
their triggering promoted by the over steepness of the scarps (Fig. 3).
In this case landsliding and faulting are mechanically related (faulting
control sites of preferred landsliding) although the landsliding postdates the faulting; (b) In 5 cases, landslide scars are located upslope of
faults, which are buried by material derived from the landslides
(Fig. 13a). Here, landsliding postdates the faulting but is not mechanically affected by it; (c) In 15 cases landslide deposits are cut by fault
scarps. In this case faulting postdates the landsliding (Fig. 13b). Finally,
in 9 cases, more than one of the above cross cutting relations exists per a
single landslide.
4. Discussion
In this work we mapped and analyzed for the first time the 447 submarine landslides and numerous fault scars exposed on the surface of
the continental slope of eastern Mediterranean, off the coast of Israel.
The landslide sizes (surface area) range between 0.0024 and 91 km2.
This size range of submarine landslides is studied here for the first
time and is complementary to previous studies of submarine landslide
population around the Mediterranean (Urgeles and Camerlenghi,
2013) and elsewhere (e.g., N. Atlantic) which analyzed larger landslides
108
O. Katz et al. / Marine Geology 369 (2015) 100–115
Fig. 7. A series of three complex landslides (For location see Fig. 2a), marked I, II, III. (a) Map view showing bathymetry. Each color set (red to blue to purple) is 200 m depth; (b) Map view,
same extent as in a, showing slope angle; (c) Perspective view; (d) Inset of c, the Primary, Secondary and Tertiary scars are marked by a continuous with P, dashed with S and dasheddotted with T lines, respectively. Arrows marks the sites where the secondary sliding material cut through the deposit of the primary landslide.
only, with areas usually N 1 km2 (Micallef et al., 2008; Chaytor et al.,
2009).
4.1. Landslide size distribution
Urgeles and Camerlenghi (2013) analyzed a population of Mediterranean landslides with area in most cases exceeding 1 km2 (their
Fig. 4). Our present analysis partly overlaps their size range but also
adds smaller landslide sizes, ranging between 10−3–101, thus extending the lower end of the size distribution studied by Urgeles and
Camerlenghi (2013), where a flattening out is observed in their data.
The power law exponent that fits the larger landslides size distribution
here (Fig. 10a) is similar to the one found in Urgeles and Camerlenghi
(2013). Thus, we suggest that the data set of Urgeles and Camerlenghi
(2013) is not complete at the small end of landslides sizes and that
the current analysis is complementary to their work, together covering
more than 9 order of area size magnitude.
Once the size distribution is extended to include also smaller
landslide sizes, it reveals (Fig. 10a) that the marine landslide probability distribution is best fit not by a power law, but by a double
Pareto distribution or an inverse Gamma distribution, as is also
true for sub-areal landslide size distributions (Malamud et al.,
2004). In the studied area the size distribution has a peak (roll
over), indicating that the area of the most common slide is
1.6 × 104 m2. For landslide larger than this size the distribution fits
a power law relationship, with exponent α = − 1.6. This exponent
is similar to that found in other populations of submarine slopefailures (Micallef et al., 2008 and using the data set of McAdoo
et al., 2000) and distinctively smaller than that measured from distributions of sub-aerial slope failures (Guzzetti et al., 2002; Malamud
et al., 2004). No maximal distribution (roll over) was observed in
Micallef et al. (2008), probably because only large landslides
(N4 km2 ) were analyzed there. The maximum in the distribution
found here is significant. The most likely landslide size is usually determined by material, geometrical or mechanical heterogeneity of
the slope (Katz and Aharonov, 2006; Stark and Guzzetti, 2009). The
lateral dimension of the most common landslide in our study is
~ 100 m. Using the relation between thickness (t) of a landslide and
its area, t = 2ε × √A, where ε = 0.05 ± 0.02 on land (Hovius et al.,
1997), predicts that such landslides have thickness of 10 m–20 m
(close to the value found in the few site where landslide thickness
was directly observed).
O. Katz et al. / Marine Geology 369 (2015) 100–115
109
Fig. 9. Landslide density along the studied continental slope. Density is quantified as the
number of mapped landslide per 1′ wide (~1.8 km) east–west stretch (horizontal line at
the upper left corner shows the stretch width).
steep, N θc, and indeed they are favored locations for landslide initiation in the north region. (b) Ongoing sedimentation (e.g., Schilman
et al., 2001) which results in a sequences of unconsolidated strata;
Fig. 8. The surface area of the 447 mapped landslides (a), the calculated log of the surface
area of 117 mapped Primary landslides (b) and the water depth at the scar head of all
mapped landslides (c) plotted as function of location (latitude) within the 110 km long
studied area (for landslides characteristics see Table S1). Vertical dashed-dotted line
marks the border line between the South and North provinces, at latitude 42.43°. Horizontal dashed line marks the 400 m depth.
4.2. Slope instability in the studied area
In the studied continental slope the following conditions promoting submarine slope failure prevail: (a) A slope angle that exceeds
the ‘apparent critical slope angle’, θc. We define θc as the maximum
observed slope angle that exhibits long-term stability. At the southern part of the studied area, head scars of the landslides appear at a
constant depth (ca. 350 m to 400 m) along an 80 km coast parallel
line (Fig. 2). The slope angle along this line (measured between the
landslide scars) is θc = 4–5° (Fig. 2b), and constitutes the apparent
critical angle in the studied area. Above and below this line (where
landslide scars are absent) the slope does not exceed θc, and thus
θc is the maximum angle for long-term stability. Similar apparent
critical slope angles are known elsewhere also (Masson et al., 2006
and references there in). The northern part of the studied area is
characterized by faults rupturing the sea floor. Faults scarps are
Fig. 10. (a) Probability density of landslide area (p) as a function of landslide area (A) for
447 landslides mapped in the studied continental slope (calculated using LStats tool by Dr.
M. Rossi, IRPI-CNR). A power law (p ~ Aα) with α = −1.6 is fitted to the large-area part
(A = 1 × 105–1 × 107 m2) of the landslide population (marked by dashed line);
(b) Cumulative area distribution of the same landslide population as in (a). A power law
of θ = − 0.7 is fitted to the large-area part (A = 1 × 105–1 × 107) of the landslide population (marked by solid line).
110
O. Katz et al. / Marine Geology 369 (2015) 100–115
Fig. 11. A seismic line (from ‘Yam Hadera’ 3D survey by Modiin Energy) crosses the northern part of the studied area (for location see Fig. 2a) and revealing the abundance of faults above
the Messinian evaporites, some of them rapturing the surface creating fault scarps, which are detectible on the sea floor using the bathymetric grid (shown above the section as an inset
from Fig. 2a).
(c) Overall slope-normal extension which is a consequence of
the down slope creeping of the over than one kilometer thick
Pliocene–Quaternary sequence on top of the Messinian evaporates
(Garfunkel, 1984). This extension is revealed by the numerous
slope-parallel growth fault documented in the studied continental
slope (e.g., Garfunkel, 1984 and Figs. 2, 11, 12); (d) Seismic activity
from multiple sources: Dead Sea Transform, Carmel fault, Cyprus
Arc and also within the studied area. Though currently there are no
direct reports of earthquake-induced submarine landslides from
the studied area, indirect interpretation of historical tsunamis as a
result of earthquake induced submarine landslides are suggested
by Salamon et al. (2007). Spatially, the northern part of the studied
area is closer to all the above mentioned seismic sources;
(e) Apparent existence of gas close to the surface (Coleman and
Ballard, 2001; Schattner et al., 2012) which might trigger sliding by
lowering sediment coherence, or by elevating pore pressure and
seepage forces (Brothers et al., 2014).
Considering the above, it is not surprising that the bathymetry of the
studied continental slope reveals an extensive mass wasting activity,
apparently taking place over long time periods. The age of the landslides
is currently unknown, but the small landslides are relatively young or
otherwise they would have been covered by the ongoing sedimentation
(see discussion below). In addition to landslides, submarine fault scarps
rupturing the surface of the continental slope were also observed in the
studied area. Their surface expression suggests that these are recent or
still active. The observed close association of the landslides and the
faults suggests that the landslides are recent or still active as well.
4.3. Observations on the slope mechanics and the structure of the slope
Landslides along the studied continental slopes in eastern Mediterranean appear over five size (area) scales (10−3–101 km2). In addition
two mega landslides are observed in the studied area, the Dor and
Palmahin disturbances (Fig. 1). The areas of these mega-structures
exceed 100 km2, and their depth involves the entire post Messinian sediment sequence (Almagor and Garfunkel, 1979; Garfunkel et al., 1979;
Almagor, 1984). The Dor and Palmahin disturbances were previously
studied, and they are not the focus of the current work and are not further discussed.
The landslides analyzed and discussed in the current work can be
separated into two groups based on their scar location: The first group
consists of landslides that originate from, or are in a close spatial association with, the relatively steep fault scarps (Fig. 3). These landslides are
typical of the northern part of studied area, north of 32.43°N, and are
relatively small. The second group, south of 32.43°N, consists of landslides with no relation to fault scarps (Fig. 4). These landslides are larger
than the ones in the former group, show characteristics of disrupted
sliding, and are not associated with faults (Figs. 2, 6). The difference
between the northern and southern landslides may be due to various
reasons:
A) In general, the surface area of landslides (slumps) in a homogeneous slope will approach the system size (Katz and
Aharonov, 2006), and the landslides' thickness will depend
on its area (e.g., Hovius et al., 1997; Klar et al., 2011). An
O. Katz et al. / Marine Geology 369 (2015) 100–115
111
Fig. 12. A seismic line (Line 2061-IS by TGS-NOPEC Geophysical Company L.P.) crosses the Dor disturbance (for location see Fig. 2a) and revealing faults above the Messinian evaporites,
some of them rapturing the surface creating antithetic fault scarps, which are detectible on the sea floor using the bathymetric grid (shown above the section as an inset from Fig. 2a).
important change in continental slope geometry occurs between its southern and northern parts. In the north of the
studied area numerous faults rupture the surface and limit
the actual possible system size, which might be the reason
for the relatively smaller size of the landslides in the northern
part (Fig. 8). In the south, the slope is not dissected by faults
(Fig. 2). Thus, un-dissected by faults, the available system
size is larger and consequently the landslides area as well.
B) Proximity to earthquake sources in the northern part of the studied area may affect landslide sizes: shorter reoccurrence time of
seismic accelerations exceeding critical acceleration for slope instability is expected there. The shorter interval between shaking
events limits the amount of sediment that accumulates between
sliding events, and makes larger landslides unlikely. Camerlenghi
et al. (2010) reviewed the important, yet surprising, observation
that large landslides are common along seismically inactive Mediterranean continental margins, and that currently fewer large
slope failures occur along tectonically active margins (e.g., the Oregon margin, McAdoo and Watts, 2004) than along passive continental margins (e.g., ten Brink et al., 2006). Thus, the larger
landslides observed in the south may partially be attributed to
the lesser seismic shaking experienced there.
C) The observed larger landslides in the south might also be partially
attributed to the larger abundance of sediments there as a results
of the proximity to the Nile outlet.
D) Another difference between landslides of the southern and northern parts of the studied continental slope is the type of landslides.
In the south, landslides are disrupted, so that material originating
within the scar flows down slope, forming lobes of deposited material, which are not fully constrained between side-scars (Fig. 4).
Slump-like surface features are more typical to the north where
in most cases deposited material is constrained between scars,
and also does not travel far down slope. A possible reason for this
difference might have been more coherent slope material in the
north, though this suggestion is not supported by recent observations showing that grain size of slope material at the surface is similar for a given water depth along the studied continental slope
(Elyashiv et al., 2014). Alternatively, the difference might be a result of the different landslide sizes, where larger landslides with
higher remobilized material volumes undergo longer runout
(Edgers and Karlsrud, 1982; Goren and Aharonov, 2007, 2008),
and are more easily fluidized in comparison to smaller landslides.
4.4. Critical slope angle
Figs. 2b and 3 show that landslides across the studied area are associated with a locally steep slope gradient. In the northern part of the
studied area, as described above, landslides are triggered from the
N10° steep fault scarps, beginning at water depth of less than 200 m.
In the southern part, a local steepening of the slope from less than
θc = 4°–5°, to over θc takes place at water depth of ca. 350 m–400 m
(Figs. 2b, 4). The majority of the landslide scars in the southern part
are situated within this steep stretch. Hence we suggest that a major
112
O. Katz et al. / Marine Geology 369 (2015) 100–115
factor affecting slope instability in the studied area is steep slope gradients. It controls the location of the landslides and apparently brings the
submarine slopes close to instability, where they can be triggered with
further change of loading conditions.
It appears that the surface expression of the two large disturbances
in the studied area, Palmahim and Dor, interacts with the smaller
scale landslides and affect their location. The Dor disturbance changes
the morphology of the shelf break, east of the disturbance the shelf is
the narrowest and faults appear at shallower depth than further south
(Fig. 2). Thus θc is achieved at shallower water depth east of the Dor disturbance. Within the perimeter of Palmahim disturbance slope angle
exceeding θc is rarely found (Fig. 2b) and indeed only one landslide
was mapped there.
The apparent critical slope angle, θc, is much lower than friction angles found by laboratory rock-mechanical tests on slope material (15°–
17° for our study area, based on consolidated-undrained triaxial compression tests; Almagor and Wiseman, 1982). The reason behind the
difference between the long-term stable slope angle θc and friction angles from lab tests is not clear, though submarine slope instability at 5°
or less are known from many other sites around the globe where continental slopes present very low slope angles (Masson et al., 2006 and references there in). Various hypotheses have been proposed to explain
the lower angle of the submarine slopes, e.g., episodes of pore pressure
elevation, retrograde sliding that brings the slope to be lower on average (Utili and Crosta, 2011) and more. This subject is beyond the
scope of the current paper.
In addition to studying the undisturbed slope angle, θc, we measured
slope angle of landslides, scars and deposits, finding three different
characteristic slope-angles: ca. 13° for head scars, ca. 5° for landslide deposits close to the scar and very low angles (1°) for the toes of landslides
(Table 1). The lowest angle, 1°, measured for landslide's toes, might be
the result of liquefaction and the dynamic angle of repose for flowing
granular material (Lajeunesse et al., 2004). At the other extreme, the
head scars are steeper than θc, which is not trivial to understand. A possible explanation follows from the fact that scar angles are similar to the
friction angle. This similarity in angles suggests that following sedimentation of the disintegrated unconsolidated slope material at θc, which
reflects a lower frictional angle, the slope material undergoes compaction, consolidation and healing processes that increase its strength toward the laboratory peak strength. Alternatively, the steep slopes of
the scars might represent a quasi- stable state, which with time will
go through a sequence of secondary failures, eventually reaching θc
(see Katz et al., 2014, Fig. 4a,b for snapshots from modeling of such a
failure sequence). A similar mechanism, with an over-steepened landslide scar reaching stability at the angle of repose via upslope
retrogressive failure, was suggested analytically by Utili (2005) and
Utili and Crosta (2011).
4.5. Possible triggering mechanisms for landsliding
Submarine landslides on the critical or supercritical slopes may be
triggered by seismic loading or loading by fluid pressure (Masson
et al., 2006). Analyzing which are the triggering mechanisms relevant
to the landslides in the studied area is the key to understanding whether
the continental slope off the coast of Israel is currently unstable and thus
whether landsliding is currently expected. To accomplish this, we integrate our newly acquired mapping and related visual information to
constrain the triggering mechanisms relevant to the studied area.
The studied area is situated at distance of 80–120 km away from the
Dead Sea fault zone (Fig. 1), which is capable of producing M N 7 earthquakes (Hamiel et al., 2009). In addition, the offshore section of the Carmel fault line is situated at the northern end of the studied area. This
fault line is believed to be capable of producing up to M 6.5 earthquakes
(Hofstetter et al., 1996). Some weak seismic activity (M b 4) is also monitored within the studied area (http://seis.gii.co.il/en/earthquake/
searchEQ.php), although the source of this activity is not comprehensively clear. Thus the studied area is subject to seismic accelerations,
which according to simple attenuation equations (Boore et al., 1997),
might be as high as 0.05 g at the southern and northern ends of the studied area (for Mw = 7 Dead Sea Transform earthquake; not including site
effects). Accordingly, previous studies, based on slope stability analysis,
show that seismic loading is a relevant triggering mechanism for the
studied area (Almagor and Wiseman, 1977, 1982; Frydman and
Talesnick, 1988). Other studies relate historical reports of earthquakeinduced tsunami, affecting the shore along the studied area, to earthquake induced submarine landsliding (Salamon et al., 2007). In summary, seismic triggering of landslide is relevant to the studied area.
Another possible triggering agent, in addition to earthquakes, is gas
seeps. Gas-seeps are often observed in close spatial proximity to underwater landslides, and were suggested to initiate some of the largest
underwater landslides on earth (Driscoll et al., 2000; Hovland et al.,
2002 and references therein). It should be noted that earthquakes and
seeps have been suggested to combine in triggering landslides
(Hovland et al., 2002). Recently it was shown that the Israeli offshore
is rich with gas-seeps (Lazar et al., 2012; Schattner et al., 2012), so
that this can be potentially an important trigger in the studied region,
though higher resolution studies must be done to understand the fluid
feature location with respect to mapped landslides (as in e.g., Hovland
et al., 2002, Fig. 4).
Fig. 13. Perspective view showing the cross-cutting relations between landslides and fault scarps (for location see Fig. 5). (a) Landslide deposit (marked by dashed line) covers and thus
postdates a 40 m high fault scarp (marked by an arrow); (b) Landslide deposit (marked by dashed line) is cut and thus predates by a 40 m high fault scarp.
O. Katz et al. / Marine Geology 369 (2015) 100–115
4.6. Timing of landsliding
Much work has been attempted around the world to constrain the
time of occurrence of submarine landslides (Lee, 2009 and reference
there in; Urlaub et al., 2013), either directly estimated using the numerical age of the pre or post sliding sediments (Normark et al., 2004) or by
using the thickness of the post sliding sediment and the estimated
sedimentation rates (Prior et al., 1986).
The timing of the studied landslides, and the question whether they
are the result of a single short event or the result of a continuous, and
possibly still active, process, are important open questions. This section
provides a first estimate for a general age constraint, using the relief of
the small submarine landslides in the studied area and the local sedimentation rates: The thickness (t) of a landslide with a mapped area
(A) of less than 0.1 km2 (242 such landslides were mapped, Fig. 2) is
expected not to exceed 30 m (t = 2ε × √A, where ε = 0.05 ± 0.02 on
land; Hovius et al., 1997. ε is probably less under water). Mechanical
constraints on the geometry of landslides do not allow the head scar
to exceed significantly the thickness. Sedimentation rates calculated
for the studied area range between a few tens centimeters per
1000 years (e.g., Hamann et al., 2008) and up to a meter per
1000 years (Schilman et al., 2001). Using sedimentation rate of 0.6 ±
0.3 m per 1000 years, we calculate that a scar relief of 30 m will be
half filled by sedimentation in 33,000 ± 16,000 years, at which point
we won't be able to resolve it anymore with our current vertical resolution. Thus, we suggest that the small landslides mapped in the studied
area, as well as large landslides which are genetically related to secondary small landslides typically triggered in their steep scars, are less than
50,000 years old assuming the slowest sedimentation rates, and are
constrained to be younger than 17,000 old if we assume sedimentation
occurs rapidly at 0.9 m per 1000 years. This time window overlaps partially with the last glacial period and also covers the transition to the
current interglacial, which is the last major global climatic change
(Torfstein et al., 2013). These age constraint also overlaps with landslide
age estimates presented in Urgeles and Camerlenghi (2013), who show
that many submarine landslides from around the Mediterranean have
occurred in the last 40,000 years (their analysis did not include the landslides mapped in the frame of this work) possibly due to climateinduced stress changes during the last deglaciation phase. These age
constraints also don't contradict a general decrease in submarine landslide activity within the Holocene (Lee, 2009 and reference therein).
To get a general idea about the question whether the close to 450
landslides mapped in the studied area represent a short or a continuous
event in time, two kinds of observations are considered: First, we
noticed that 57 landslides in the northern part of the open slope are triggered from the over steepened fault scarps (e.g., Fig. 3; Table S1). Hence,
landsliding and faulting are mechanically as well as generally temporally related. The geological evidence (Garfunkel, 1984) suggest that
faulting within the Pliocene–Pleistocene sequence along the continental
slope is syn-sedimentary, temporally-continuous, and apparently still
active, post-dating the Messinian time and under the overall extension
resulting from the basin-ward creeping of this sequence on the
Messinian evaporates (Frey-Martinez et al., 2005). Hence, if faulting is
continuous over geological time scale and possibly still on-going, landslides are also expected to behave similarly, as the latter are promoted
by surface rupturing and steepening of the faults. These relations also
suggest that the dynamics of the Messinian evaporites at depth is a possible drive for at least some of the studied shallow landslides; Additional
evidence for long and continuous morphological activity in the studied
area is the complicated crosscutting relation between landslides and
fault scarps (Fig. 13) and between different generations of landslides
(Fig. 5). This kind of morphology is highly unlikely to develop instantaneously, rather it is expected to develop in a long and continuously sequence of events.
Based on the observed structures and morphology, we thus suggest
a general time frame for the sliding activity in the studied area. The
113
small and large submarine landslides occurred continuously at least
throughout the last few tens of thousands of years and are still occurring. Based on outside evidence from elsewhere in the Mediterranean
(Urgeles and Camerlenghi, 2013), it is likely that sliding events increased in number toward the end of the last glacial period and into
the current interglacial. Our time frame does not contradict the one of
Urgeles and Camerlenghi (2013), though numerical dating methods
are needed to know whether the processes and timing suggested in
Urgeles and Camerlenghi (2013), are fully applicable to the studied
landslides.
4.7. Submarine landslide hazard
The likelihood of tsunami generation, and the severity of direct impact damage, increases directly with landslide size. Hence in general,
open continental slope landslides, which tend to be the largest, deserve
the greatest attention in any submarine landslide hazard evaluation
(Lee, 2009). Such logic places the studied area in high hazard relative
to the canyon province to its north (Fig. 1). Within the studied area,
landslide sizes decrease northwards (Fig. 8) and hence the hazard is
somewhat lower in the north, considering landslide size only and assuming that future landslides will be generally similar to the mapped
one. On the other hand, considering the landslide water depth and distance from the shoreline, landslides in the northern part of the studied
area occur at shallower water depths (Fig. 8) and closer to the shore
(Fig. 2).
Temporally, the interglacial period is considered the most stable one
in the glaciation de-glaciation Pleistocene–Holocene cycles (Lee, 2009)
thus we expect a decrease in landslide activity in the last 5000 years
or so. Our preliminary qualitative assessment of the landslide ages
does not contradict, nor does it support, this observation. Although occurrence of landslide induced tsunamis in the last centuries were suggested based on historical documents (Salamon et al., 2007), we could
not highlight a particular landslide that presents an outstanding distinctive recent morphology. Several landslide characteristics affect tsunami
generation, where tsunami hazard increases with the thickness of the
landslide, its area, and its propagation velocity (Ward, 2001). Rotational
landslides or slumps, which consist of a thick slide block and a steep
headwall, if they move rapidly enough down-slope, are potentially
extremely hazardous for generating tsunamis, even when the lateral
distance moved is small and little effect is seen on the seafloor downslope of the immediate landslide site (Masson et al., 2006). The Papua
New Guinea 1998 tsunami, which killed N2000 people, was likely generated in this way (Matsumoto and Tappin, 2003; Sweet and Silver,
2003). Our studied area shows quite a few steep cohesive slumps. Further study using our observed landslide dimensions and simulating
their motion, should be used to assess their tsunamogenic hazard. For
that purpose, landslide volumes in the studied area can be directly estimated from our measured landslides area using published landslide
area–volume scales (e.g., ten Brink et al., 2006; Klar et al., 2011). As a
first approximation for the tsunami hazard, calculating the landslide
volumes (V) using V = 0.0267A1.032 deduced for clay rich submarine
landslides (ten Brink et al., 2006), reveals that only 3 of the 447 mapped
submarine landslides have V larger than 1 km3, the minimal size of producing noticeable tsunami (ten Brink et al., 2006). These are located at
the southern part of the studied area.
If a seismic trigger is considered for the submarine landslides in the
studied area (e.g., Frydman and Talesnick, 1988), the sliding events
reoccurrence time has to generally follow the one of the earthquakes.
Only large enough earthquakes will induce submarine slope failure in
the studied area. Setting a threshold magnitude of 7.0 (for a Dead Sea
transform earthquake, following Keefer, 1984), a reoccurrence time
larger than 1000 years is expected (Hamiel et al., 2009). If the trigger
is different than seismic, a different, yet unavailable, reoccurrence
time is applicable. Reoccurrence time of ca. 100 years for landslides
with volume larger than 0.1 km3 is suggested by Urgeles and
114
O. Katz et al. / Marine Geology 369 (2015) 100–115
Camerlenghi (2013) for the Mediterranean. These landslides are capable of producing damage to coastal and offshore facilities but are not
considered tsunamogenic. Considering the results of the current work,
tens such landslides are mapped (calculating volumes using the V–A relation used above) within a general time frame of tens thousand years,
i.e., calculated mean reoccurrence time of once every thousands years.
5. Conclusions
We mapped and analyzed for the first time the 447 small to medium
submarine landslides and numerous fault scars exposed on the surface
of the continental slope of eastern Mediterranean, off the coast of
Israel. Failed slopes cover more than 20% of the studied continental
slope. Landslide sizes (surface area) range between 0.0024 km2 and
91 km2, where the observed size distribution has a peak (roll over)
showing that the most probable landslide area is 1.6 × 104 m2. For landslide larger than this size the distribution fits a power law relationship,
with exponent α = −1.6, which is similar to that found in other populations of submarine slope-failures. This size range of submarine landslides is studied here for the first time and is complementary to
previous studies around the Mediterranean and elsewhere which analyzed larger landslides only. Nevertheless the Landslides are not evenly
distributed along the studied continental slope, in the northern part
they are generally smaller and found in shallower water depth than
the ones in the south. The morphology of the landslides as well as
their cross cutting relation with the fault scarps suggest that these landslide are recent, apparently younger than 50,000 year, and possibly still
active. The triggering mechanism is not clear yet but a few processes are
suggested as promoting slope instability in the studied area: (a) slopes
in the studied area are close to a critical slope angle which is observed
to be 4°–5°; (b) Continuous sedimentation which increases the load
on the slope; (c) Active salt tectonic which results in an overall extension and surface rupturing and steepening by normal faults;
(d) Seismic loading; and (e) Apparent existence of gas close to the surface. These observations highlight the possible current geo-hazard imposed by the landslides and call for further research.
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.margeo.2015.08.006.
Acknowledgments
This work was funded by the Ministry of Science and Technology,
Israel (grant number 3–9145). Bathymetric grids were kindly made
available to us by Dr. John K. Hall (Geological Survey of Israel). We
thank Dr. Mauro Rossi (IRPI-CNR, Italy) for his generous help in calculating the size distribution of the landslides (Fig. 10a) using LStats tool developed in the frame of FP7-LAMPRE project.
References
Almagor, G., 1984. Salt-controlled slumping on the Mediterranean slope of central Israel.
Mar. Geophys. Res. Lett. 6, 227–243.
Almagor, G., Garfunkel, Z., 1979. Submarine slumping on the continental margin of Israel
and northern Sinai. Bull. Am. Assoc. Pet. Geol. 63, 324–340.
Almagor, G., Wiseman, G., 1977. Analysis of submarine slumping on the continental slope
off the southern coast of Israel. Mar. Geotechnol. 2, 349–388.
Almagor, G., Wiseman, G., 1982. Submarine Slumping and Mass Movements on the Continental Slope of Israel. Marine Slides and Other Mass Movements NATO Conference
Series 6, pp. 95–128.
Almogi-Labin, A., Bar-Matthews, M., Shriki, D., Kolosovsky, E., Paterne, M., 2009. Climatic
variability during the last w90 ka of the southern and northern Levantine Basin as evident from marine records and speleothems. Quat. Sci. Rev. 28, 1–15.
Baudon, C., Cartwright, J., 2008. The kinematics of reactivation of normal faults using high
resolution throw mapping. J. Struct. Geol. 30, 1072–1084.
Ben-Avraham, Z., 1978. The structure and tectonic setting of the Levant continental margin, Eastern Mediterranean. Tectonophysics 46, 313–331.
Ben-Gai, Y., Ben-Avraham, Z., Buchbinder, B., Kendall, C.G.St.C., 2005. Post-Messinian evolution of the Southeastern Levant Basin based on two-dimensional stratigraphic simulation. Mar. Geol. 221, 359–379.
Boore, D.M., Joyner, W.B., Fumal, T.E., 1997. Equations for estimating horizontal response
spectra and peak acceleration from western North American earthquakes: a summary of recent work. Seismol. Res. Lett. 68, 127–153.
Booth, J.S., O'Leary, D.W., Popenoe, P., Danforth, W.W., 1993. U.S. Atlantic continental
slope landslides: their distribution, general attributes, and implifications. Submarine
landslides: Selected Studies in the U.S. exclusive economic zone. U.S. geological survey Bulletin, pp. 14–22.
Brenner, S., 2003. High-resolution nested model simulations of the climatological circulation in the southeastern. Mediterr. Sea Ann. Geophys. 21, 267–280.
Brothers, D.S., Ruppel, C., Kluesner, J.W., Brink, U.S., Chaytor, J.D., Hill, J.C., Andrews, B.D.,
Flores, C., 2014. Seabed fluid expulsion along the upper slope and outer shelf of the
US Atlantic continental margin. Geophys. Res. Lett. 41, 96–101.
Bull, S., Cartwright, J., Huuse, M., 2009. A subsurface evacuation model for submarine
slope failure. Basin Res. 21, 433–443.
Calvert, S.E., Fontugne, M.R., 2001. On the late Pleistocene–Holocene sapropel record of
climatic and oceanographic variability in the eastern Mediterranean.
Paleoceanography 16, 78–94.
Camerlenghi, A., Urgeles, R., Fantoni, L., 2010. A Database on Submarine Landslides of the
Mediterranean Sea. In: Mosher, C.D., et al. (Eds.), Submarine mass movements and
their consequences: advances in natural and technological hazards research vol. 28,
III. Springer Science + Business Media B.V., pp. 503–514.
Cartwright, J.A., Jackson, M.P.A., 2008. Initiation of gravitational collapse of an evaporate
basin margin: the Messinian saline giant, Levant Basin, eastern Mediterranean. GSA
Bull. 120 (3/4), 399–413.
Cartwright, J., Jackson, M., Dooley, T., Higgins, S., 2012. Strain partitioning in gravity-driven shortening of a thick, multilayered evaporite sequence. In: Alsop, G.I., Archer, S.G.,
Hartley, A.J., Grant, N.T., Hodgkinson, R. (Eds.), Salt Tectonics, Sediments and
Prospectivity. Geological Society, London, Special Publications, pp. 449–470.
Chaytor, J.D., ten Brink, U.S., Solow, A.R., Andrews, B.D., 2009. Size distribution of submarine landslides along the U.S. Atlantic margin. Mar. Geol. 264, 16–27.
Coleman, D.F., Ballard, R.D., 2001. A highly concentrated region of cold hydrocarbon seeps
in the southeastern Mediterranean sea. Geo-Mar. Lett. 21, 162–167.
Driscoll, N.W., Weissel, J.K., Goff, J.A., 2000. Potential for large-scale submarine slope
failure and tsunami generation along the U.S. mid-Atlantic coast. Geology 28,
407–410.
Edgers, L., Karlsrud, K., 1982. Soil flows generated by submarine slides: case studies and
consequences. Norw. Geotech. Inst. Bull. 143, 1–11.
Elyashiv, H., Crouvi, O., Almogi-Labin, A., HarLavan, Y., Hyams-Kaphzan, O., 2014. Characteristics of deep sea sediments from the Levantine basin (Israel economic
zone)—preliminary results. 19th International Sedimentological Congress, Geneva,
Switzerland.
Frey-Martinez, J., 2010. 3D seismic interpretation of mass transport deposits: implication
for basin analysis and geohazard evaluation. In: Mosher, et al. (Eds.), Submarine Mass
Movements and Their ConsequencesAdvances in Natural and Technological Hazards
Research 28. Springer Science + Business Media B.V., pp. 553–568.
Frey-Martinez, J., Cartwright, J., Hall, B., 2005. 3D seismic interpretation of slump complexes: examples from the continental margin of Israel. Basin Res. 17, 83–108.
Frydman, S., Talesnick, M.L., 1988. Analysis of seismically triggered slides of Israel. Environ. Geol. 11, 21–26.
Garfunkel, Z., 1984. Large-scale submarine rotational slumps and growth faults in the
eastern Mediterranean. Mar. Geol. 55, 305–324.
Garfunkel, Z., Arad, A., Bugge, T., Almagor, G., 1979. The Palmahim Disturbance and its regional setting. Geol. Surv. Israel Bull. 72.
Goren, L., Aharonov, E., 2007. Long runout landslides: the role of frictional heating and hydraulic diffusivity. Geophys. Res. Lett. 4, L07301. http://dx.doi.org/10.1029/
2006GL028895.
Goren, L., Aharonov, E., 2008. On the stability of landslides: a thermo-poro-elastic approach. Earth Planet. Sci. Lett. http://dx.doi.org/10.1016/j.epsl.2008.11.002.
Gradmann, S., Hubscher, C., Ben-Avraham, Z., Gajewski, D., Netzband, G., 2005. Salt tectonics offnorthern Israel. Mar. Pet. Geol. 22, 597–611.
Guzzetti, F., Malamud, B.D., Turcotte, D.L., Reichenbach, P., 2002. Power-law correlations
of landslide areas in central Italy. Earth Planet. Sci. Lett. 195, 169–183.
Gvirtzman, G., Buchbinder, B., 1978. The Late Tertiary of the coastal plain and continental
shelf of Israel and its bearing on the history of the eastern Mediterranean. In: Ross,
D.A., et al. (Eds.), Initial Rep. of the Deep Sea Drilling Project vol. 42A. U.S. Government Printing, Washington, pp. 1195–1222.
Gvirtzman, Z., Reshef, M., Buch-Leviatan, O., Groves-Gidney, G., Karcz, Z., Makovsky, Y.,
Ben-Avraham, Z., 2015. Bathymetry of the Levant basin: interaction of salt-tectonics
and surficial mass movements. Mar. Geol. 360, 25–39.
Hamann, Y., Ehrmann, W., Schmiedl, G., Kruger, S., Stuut, J.B., Kuhnt, T., 2008. Sedimentation processes in the Eastern Mediterranean Sea during the Late Glacial and Holocene
revealed by end-member modelling of the terrigenous fraction in marine sediments.
Mar. Geol. 248, 97–114.
Hamiel, Y., Amit, R., Begin, Z.B., Marco, S., Katz, O., Salamon, A., Zilberman, E., Porat, N.,
2009. The seismicity along the Dead Sea fault during the last 60,000 years. Bull.
Seismol. Soc. Am. 99, 2020–2026.
Hampton, M.A., Lee, H.J., Locat, J., 1996. Submarine landslides. Rev. Geophys. 34, 33–59.
Hofstetter, A., van Eck, T., Shapira, A., 1996. Seismic activity along the fault branches of the
Dead Sea-Jordan Transform System: the Carmel–Tirtza fault system. Tectonophysics
267, 317–330.
Hovius, N., Stark, C.P., Allen, P.A., 1997. Sediment flux from a mountain belt derived by
landslide mapping. Geology 25, 231–234.
Hovland, M., Gardner, J.V., Judd, A.G., 2002. The significance of pockmarks to understanding fluid flow processes and geohazards. Geofluids 2, 127–136.
Hungr, O., Leroueil, S., Picarelli, L., 2014. The Varnes classification of landslide types, an
update. Landslides 11, 167–194.
O. Katz et al. / Marine Geology 369 (2015) 100–115
Hutton, W.H., Syvitski, P.M., 2004. Advances in the numerical modeling of sediment failure during the development of a continental margin. Mar. Geol. 203, 267–380.
Issler, D., De Blasio, F.V., Elverhoi, A., Bryn, P., Lien, R., 2005. Scaling behaviour of clay-rich
submarine debris flows. Mar. Pet. Geol. 22, 187–194.
Katz, O., Aharonov, E., 2006. Landslides in vibrating sand-box: what controls types of
slope-failure and frequency magnitude relations? Earth Planet. Sci. Lett. 247,
280–294.
Katz, O., Morgan, J.K., Aharonov, E., Dugan, B., 2014. Controls on the size and geometry of
landslides: insights from discrete element numerical simulations. Geomorphology
220, 104–113.
Keefer, D.K., 1984. Landslides caused by earthquakes. Geol. Soc. Am. Bull. 95, 406–421.
Klar, A., Aharonov, E., Kalderon Asael, B., Katz, O., 2011. Analytical and observational relations between landslide volumes and surface areas. J. Geophys. Res. 116, F02001.
http://dx.doi.org/10.1029/2009JF001604.
Korup, O., 2005. Geomorphic imprint of mass movements on alpine river systems, southwest New Zealand. Earth Surf. Process. Landf. 30, 783–800.
Kuhnt, T., Schmiedl, G., Ehrmann, W., Hamann, Y., Andersen, N., 2008. Stable isotopic
composition of Holocene benthic foraminifers from the Eastern Mediterranean Sea:
past changes in productivity and deep water oxygenation. Palaeogeogr.
Palaeoclimatol. Palaeoecol. 268, 106–115.
Lajeunesse, E., Mangeney-Castelnau, A., Vilotte, J.P., 2004. Spreading of a granular mass on
a horizontal plane. Phys. Fluids 16, 2371–2381.
Lazar, M., Schattner, U., Reshef, M., 2012. The great escape: an intra‐Messinian gas system
in the eastern Mediterranean. Geophys. Res. Lett. 39, L20309. http://dx.doi.org/10.
1029/2012GL053484.
Lee, H.J., 2009. Timing of occurrence of large submarine landslides on the Atlantic Ocean
margin. Mar. Geol. 264, 53–64.
Locat, J., Lee, H., 2002. Submarine landslides: advances and challenges. Can. Geotech. J. 39,
193–212.
Longva, O., Janbu, N., Blikra, L.H., Boe, R., 2003. The 1996 Finneidfjord slide: seafloor failure
and slide dynamics. In: Locat, J., Mienert, J. (Eds.), Submarine Mass Movements and
their Consequences. Kluwer Academic Publishers, Dordrecht, pp. 531–538.
Luz, B., Perelis-Grossowicz, L., 1980. Oxygen isotopes, biostratigraphy and recent rates of
sedimentation in the eastern Mediterranean off Israel. Isr. J. Earth Sci. 29, 140–146.
Malamud, B.D., Turcotte, D.L., Guzzetti, F., Reichenbach, P., 2004. Landslide inventories
and their statistical properties. Earth Surf. Process. Landf. 29, 687–711.
Mart, Y., Ryan, W., 2007. The Levant Slumps and the Phoenician Structures: collapse features along the continental margin of the southeastern Mediterranean Sea. Mar.
Geophys. Res. 28, 297–307.
Mart, Y., Eisin, B., Folkman, Y., 1978. The Palmahim structure—a model of continuous tectonic activity since the Upper Miocene in the Southeastern Mediterranean off Israel.
Earth Planet. Sci. Lett. 39, 328–334.
Masson, D.G., Harbitz, C.B., Wynn, R.B., Pedersen, G., Lovholt, F., 2006. Submarine landslides: processes, triggers and hazard prediction. Philos. Trans. R. Soc. A 364,
2009–2039.
Matsumoto, T., Tappin, D.R., 2003. Possible Coseismic Large-scale Landslide off the Northern Coast of Papua New Guinea in July 1998: geophysical and geological results from
SOS cruises. Pure Appl. Geophys. 160, 1923–1943.
McAdoo, B.G.L., Pratson, L.F., Orange, D.L., 2000. Submarine landslide geomorphology, US
continental slope. Mar. Geol. 169, 103–136.
McAdoo, B.G., Watts, P., 2004. Tsunami hazard from submarine landslides on the Oregon
continental slope. Mar. Geol. 203, 235–245.
Micallef, A., Berndt, C., Masson, D.G., Stow, D.A.V., 2008. Scale invariant characteristics of
the Storegga Slide and implications for large-scale submarine mass movements.
Mar. Geol. 247, 46–60.
Mulder, R., Cochonat, P., 1996. Classification of offshore mass movements. J. Sed. Res. 66,
43–57.
Normark, W.R., McGann, M., Sliter, R., 2004. Age of Palos Verdes submarine debris avalanche, southern California. Mar. Geol. 203, 247–259.
Prior, D.B., Doyle, E.H., Neurauter, T., 1986. The Currituck slide, Mid-Atlantic continental
slope—revisited. Mar. Geol. 73, 25–45.
Roter, M., 2011. High resolution survey of the Dor Submarine Slide: recent activity in the
shallow sub surface (M.Sc. Thesis), University of Haifa (77 pp., in Hebrew).
Sade, R., 2007. Morphology and acoustic backscatter of the northern Israel continental
margin based on high resolution multibeam sonar (M.Sc. Thesis), Tel Aviv University
(95 pp.).
Sade, R., Hall, J.K., Amit, G., Golan, A., Gur-Arieh, L., Tibor, G., 2007. The Israel national
bathymetric survey—a new look at the seafloor of Israel. Israel J. Earth Sci. 55,
185–187.
Salamon, A., Rockwell, T., Ward, S.N., Guidoboni, E., Comastri, A., 2007. Tsunami hazard
evaluation of the Eastern Mediterranean: historical analysis and selected modeling.
Bull. Seismol. Soc. Am. 97, 705–724.
115
Schattner, U., Lazar, M., Harari, D., Waldmann, N., 2012. Active gas migration systems offshore northern Israel, first evidence from seafloor and subsurface data. Cont. Shelf
Res. 48, 167–172.
Schilman, B., Bar-Matthews, M., Almogi-Labin, A., Luz, B., 2001. Global climate instability
reflected by Eastern Mediterranean marine records during the late Holocene.
Palaeogeogr. Palaeoclimatol. Palaeoecol. 176, 157–176.
Shaller, P.J., Smith-Shaller, A., 1996. Review of proposed mechanisms for sturzstroms
(long-runout landslides). In: Abbott, P.L., Seymour, D.C. (Eds.), Sturzstroms and Detachment Faults. S. Coast Geol. Soc., Santa Ana, Calif., pp. 185–202.
Stanley, D.J., Nir, Y., Galili, E., 1998. Clay mineral distributions to interpret Nile cell provenance and dispersal: III. Offshore margin between Nile Delta and Northern Israel.
J. Coast. Res. 14, 196–217.
Stark, C.P., Guzzetti, F., 2009. Landslide rupture and the probability distribution of mobilized debris volumes. J. Geophys. Res. 114, F00A02. http://dx.doi.org/10.1029/
2008JF001008.
Sultan, N., Cochonat, P., Canals, M., Cattaneo, A., Dennielou, B., Haflidason, H., Laberg, J.S.,
Long, D., Mienert, J., Trincardi, F., Urgeles, R., Vorren, T.O., Wilson, C., 2004. Triggering
mechanisms of slope instability processes and sediment failure on continental
margins: a geotechnical approach. Mar. Geol. 213, 291–321.
Sweet, S., Silver, E.A., 2003. Tectonics and slumping in the source region of the 1998 Papua
New Guinea tsunami from seismic reflection images. Pure Appl. Geophys. 160,
1945–1968.
Tappin, D.R., 2010. Submarine mass failures as tsunami sources: their climate control.
Philos. Trans. R. Soc. A 368, 2417–2434.
ten Brink, U.S., Geist, E.L., Andrews, B.D., 2006. Size distribution of submarine landslides
and its implication to tsunami hazard in Puerto Rico. Geophys. Res. Lett. 33,
L11307. http://dx.doi.org/10.1029/2006GL026125.
Thomas, S., Hooper, J., Clare, M., 2010. Constraining geohazards to the past: impact assessment of submarine mass movements on seabed developments. In: Mosher, C.D.,
Shipp, R.C., Moscardelli, L., Chaytor, J.D., Baxter, C.D.P., Lee, H.J., Urgeles, R. (Eds.), Submarine Mass Movements and Their Consequences: Advances in Natural and Technological Hazards Research vol. 28, II, pp. 387–398.
Tibor, G., Ben-Avraham, Z., Steckler, M., Fligelman, H., 1992. Late Tertiary subsidence
history of the southern Levant Margin, eastern Mediterranean Sea, and its implications to the understanding of the Messinian event. J. Geophys. Res. 97,
17,593–17,614.
Tibor, G., Sade, R., Sade, H., Hall, J.K., 2013. Data collection and processing of multibeam
data from the deep water offshore Israel. IOLR report H-31/2013.
Tinti, S., Armigliato, A., Manucci, A., Pagnoni, G., Zaniboni, F., Yalçiner, A.C., Altinok, Y.,
2006. The generating mechanisms of the August 17, 1999 İzmit bay (Turkey) tsunami: regional (tectonic) and local (mass instabilities) causes. Mar. Geol. 225, 311–330.
Torfstein, A., Goldstein, S.L., Stein, M., Enzel, Y., 2013. Impacts of abrupt climate changes in
the Levant from Last Glacial Dead Sea levels. Quat. Sci. Rev. 69, 1–7.
Tripsanas, E.K., Piper, D.J.W., Jenner, K.A., Bryant, W.R., 2008. Submarine mass-transport
facies: new perspectives on flow processes from cores on the eastern North
American margin. Sedimentology 55, 97–136.
Twichell, D.C., Chaytor, J.D., ten Brink, U.S., Buczkowski, B., 2009. Morphology of the late
Quaternary submarine landslides along the U.S. Atlantic continental margin. Mar.
Geol. 264, 4–15.
Urgeles, R., Camerlenghi, A., 2013. Submarine landslides of the Mediterranean Sea: trigger
mechanisms, dynamics, and frequency–magnitude distribution. J. Geophys. Res. 118,
2600–2618.
Urgeles, R., Locat, J., Sawyer, D.E., Flemings, P.B., Dugan, B., Binh, N.T.T., 2010. History of
pore pressure build up and slope instability in mud-dominated sediments of Ursa
Basin, Gulf of Mexico Continental Slope. In: Mosher, et al. (Eds.), Submarine Mass
Movements and Their ConsequencesAdvances in Natural and Technological Hazards
Research 28. Springer Science + Business Media B.V., pp. 179–190.
Urlaub, M., Talling, P.J., Masson, D.G., 2013. Timing and frequency of large submarine
landslides: implications for understanding triggers and future geohazard. Quat. Sci.
Rev. 72, 63–82.
Utili, S., 2005. An analytical relationship for weathering induced slope retrogression: a
benchmark. Riv. Ital. Geotech. 39, 9–30.
Utili, S., Crosta, G.B., 2011. Modeling the evolution of natural cliffs subject to weathering:
1. Limit analysis approach. J. Geophys. Res. 116. http://dx.doi.org/10.1029/
2009JF001557.
Ward, S.N., 2001. Landslide tsunami. J. Geophys. Res. 106, 11201–11215.
Winkelmann, D., Geissler, W.H., Stein, R., Niessen, F., 2010. Post-megaslide slope stability
north of Svalbard, Arctic Ocean. In: Mosher, et al. (Eds.), Submarine Mass Movements
and Their ConsequencesAdvances in Natural and Technological Hazards Research 28.
Springer Science + Business Media B.V., pp. 279–287.