Document Actions

ijege-13_bs-mazzanti-de-blasio.pdf

background image
Italian Journal of Engineering Geology and Environment - Book Series (6) www.ijege.uniroma1.it © 2013 Sapienza Università
Editrice
425
DOI: 10.4408/IJEGE.2013-06.B-41
LANDSLIDES SURFING ON WATER: A PRELIMINARY STUDY
P
aolo
MAZZANTI
(*,**)
& F
abio
V
ittorio
DE BLASIO
(*,***)
(*)
University of Rome “Sapienza - Department of Earth Sciences, P.le Aldo Moro 5 - 00185 Rome, Italy
Ph: 0649914835 - email: paolo.mazzanti@uniroma1.it
(**)
NHAZCA s.r.l., spin-off University of Rome ”Sapienza”, Via Cori snc - 00177, Rome, Italy
(***)
University of Milan “Bicocca” - Department of Geosciences and Geotechnologies - Milan, Italy
lakes were recently found also in Switzerland (Lucerna
lake, K
ramer
et alii, 2012). These kinds of mixed land-
slides are also interesting in geomorphology for the
reason that the impact with water tends to re-distribute
the sediments through larger submerged areas.
The incidence of landslides that may potentially af-
fect the hill slopes surrounding lakes is increasing due to
the impoundment of several reservoirs for both water pro-
vision and hydropower energy. This is a common feature
in Alpine environment (e.g. Italy, Switzerland, Austria),
but also in several developed and developing countries
like USA, China, Russia, India, or Ethiopia. Apart from
huge disasters like the 1963 Vajont, several landslides do
occur in artificial lakes during impounding phases or also
under stationary conditions (e.g. S
chuSter
, 1979). These
problems have been recently outlined by the induced ef-
fects of a massive project like the Three Gorges Dam in
China. As a consequence of rising water level in the huge
basin, several landslides occurred along the bankside
(Y
uSheng
, 2010; b
olin
et alii, 2010), often generating
high waves. The one occurred in Gongjiafang slope on
the 23 November 2008 induced a 30 m height wave thus
causing severe economic impact (b
olin
et alii, 2012).
Several authors have attempted to understand the
main mechanism controlling the propagation of land-
slides entering a water basin and to derive information
on the ensuing tsunami propagation. Laboratory experi-
ments (F
ritz
et alii, 2003; D
i
r
iSio
et alii, 2009; m
o
-
hammeD
& F
ritz
, 2010) and back analysis of past events
(e.g., F
reunDt
et alii, 2007; m
azzanti
& b
ozzano
,
ABSTRACT
We conjecture that in some cases, landslides im-
pacting onto a water surface might acquire a vertical
momentum that makes them slide horizontally at the
water level, instead of plunging immediately into deep
water, a process that we name surfing. An example of
this behavior could be the recent (2002) landslide from
the Sciara del Fuoco (Stromboli, Italy), which caused
a tsunami with 10-15 meters high run-up waves. By
examination of photographs, laboratory experiments,
theoretical estimates, and numerical calculations, we
preliminarily investigate the surfing conditions for
landslides. The effect might also have an impact on
the generation and propagation of tsunami waves.
INTRODUCTION
It is well known that landslides impacting onto
a water basin may generate potentially destructive
tsunamis. In the last 230 years, two major tsunamis
produced by the collapse of rapid massive landslides
in water basins took place in Italy, causing more than
3500 casualties. Together with the Vajont disaster oc-
curred on 6
th
October 1963 (S
emenza
, 2002), the 1783
Scilla event can be considered one of the most destruc-
tive tsunami induced by subaerial landslides (m
azzan
-
ti
& b
ozzano
, 2011). Other similar events occurred in
the last years in natural and artificial water reservoirs
in Italy (S
emenza
, 2002) and in several other countries
(see e.g., m
azzanti
, 2008 for a review). Evidence of
ancient catastrophic tsunamis induced by landslides in
background image
P. MAZZANTI & F.V. DE BLASIO
426
International Conference Vajont 1963-2013. Thoughts and analyses after 50 years since the catastrophic landslide Padua, Italy - 8-10 October 2013
FIELD OBSERVATIONS
Some video recorded during the occurrence of
oastal landslides apparently show the surfing effect.
The most interesting has been shot by Dr. Massimo
Pompilio of INGV (Istituto Nazionale di Geofisica e
Vulcanologia), during the 29 December 2002 tsunami-
genic landslide at Stromboli volcano, Italy (c
hiocci
et
alii, 2008). Figure 2 shows some frames of the video.
The granular avalanche travelled over the sea surface
for more than 100 m in the span of few seconds, sig-
nificantly increasing its front thickness before coming
to rest. In this event, the surfing mass was only a part
of the whole landslide volume. The remaining part of
the mass probably travelled along the underwater slope
like a debris avalanche, perhaps to be transformed into
a turbidity current. Although the frames of Figure 2
show that the percentage of surfing mass was not neg-
ligible, it is difficult to give a precise estimate.
These effects have been recently observed in an-
other interesting case on the three Gorges water basin
on the Gongjiafang landslide in Wu Gorge occurred on
November 23, 2008 (b
olin
et alii, 2008). The sequence
of frames reported on the paper by b
olin
et alii (2008)
shows that a percentage of the landsliding material re-
mained at the water surface. Also in this case, the land-
slide velocity was moderate (up to 11 m/s). In spite of
the density of the rocky material involved, it appears
from the published figures that the landslide traveled a
great deal horizontally before disappearing into the wa-
ter, which is much deeper than the landslide thickness.
Thus, a certain amount of surfing must have taken place.
Other indications could be based on the obser-
vation of old landslides deposits, rather than filmed
events. For example, the chalk cliffs of Northern Eu-
rope exhibit long several tongues in correspondence
of the scars produced by slope failures on the tidal
flats (h
utchinSon
, 2002). Such mobility might have
been promoted by surfing or perhaps also by hydro-
planing on a light and impermeable medium (e.g., D
e
b
laSio
, 2011). A rather extreme case of surfing could
occur with ice avalanches collapsing in the fjords or
2011) have been performed over the last years, leading
to the development of ever-improving computer codes
aimed at back simulating events already occurred and,
potentially, to predict the tsunami impact related to fu-
ture landslides. However, it is noteworthy that the basic
mechanisms controlling the air to water transition of a
landslide are still far to be exhaustively understood, and
they have not perhaps attracted the required attention.
In a previous paper we preliminarily listed a se-
ries of simple observations apparently unnoticed or
neglected in previous investigations concerning the
impact of the mass with water and its propagation
(m
azzanti
& D
e
b
laSio
, 2011).
In this paper we focus mainly on the surfing, i.e. a
specific process that, to our knowledge, has not been
formerly investigated with reference to landslides..
By analysis of video captured during real landslides,
simple laboratory experiments and theoretical analysis
(also supported by mathematical and numerical com-
putations) we suggest that, under particular conditions,
landslides might flow almost horizontally at the inter-
face between water and air, before finally plunging into
the water (Fig. 1). Such an effect might significantly
modify the propagation of the landslide in the immedi-
ate phases that follow the impact with water, as well as
the mechanics of tsunami generation.
Preliminary considerations about the conditions
leading to the surfing process of landslides and the
potential implications on the tsunami propagation
and generation are herein presented. We will pre-
liminarily consider the general physics of impact
and surfing, while the process of tsunami generation
by surfing landslides will not be addressed in detail.
INVESTIGATING THE SURFING PRO-
CESS ON LANDSLIDES
In order to assess whether and how surfing process-
es may affect the dynamics of coastal landslides impact-
ing the water surface, in the present work we combine
observation of real landslide events, simple laboratory
experiments, and basic mathematical analysis.
Fig. 1 - Sketch showing the process of surfing as defined in the present work (time sequence from left to right). A landslide
plunging at sufficiently low impact angle against a water reservoir, lake, or the sea, may be lifted a short time, so
propagating nearly horizontally at the water-air interface
background image
LANDSLIDES SURFING ON WATER: A PRELIMINARY STUDY
Italian Journal of Engineering Geology and Environment - Book Series (6) www.ijege.uniroma1.it © 2013 Sapienza Università
Editrice
427
velocity of the tsunami wave becomes closer to the
velocity of the landslide with a distinct amplification
effect; moreover, a splash effect may create a high
train of waves at once (F
ritz
et alii, 2003a; 2003b).
We will suggest that surfing might affect the tsunami
generation for a landslide starting subaerially
LABORATORY EXPERIMENTS
Some research groups have performed experiments
simulating landslides plunging onto a water reservoir
(F
ritz
et alii, 2003, 2004; F
reunDt
, 2003; m
ohammeD
& F
ritz
, 2012). The laboratory experiments by F
reunDt
(2003) aimed at simulating the occurrence of pyroclas-
tic flows at the coastline, while the video by F
reunD
et
alii (2003) shows a bounce upon the water surface and
a surfing effect at the early stages of the water impact.
In the mentioned papers, the landslide flow has
been thoroughly investigated especially concerning the
associated tsunami, while the possibility of surfing was
the ocean (Fig. 3). Because ice is positively buoyant,
surfing will become the leading effect in this case,
even though such ice collapses are not particularly
large or worrisome. One of the most important con-
sequences of such events is the associated tsunami. It
is well-known that the Stromboli avalanche produced
a 15 m high tsunami that swept the coast of the island
(t
inti
et alii, 2005). Landslide-generated tsunamis
have been studied in detail at least since the Grand
Banks event of 1929 caused by a submarine landslide
offshore Newfoundland (F
ine
et alii, 2005). However,
landslides like the one of Grand Banks travel under-
water from start to stop. When a landslide travels com-
pletely submerged the front causes water to rise above
the unperturbed level, while the rear of the landslide
draws it downward. In other words, water is perturbed
vertically, while the shear impulse is negligible. In
contrast, a tsunami generated by a landslide starting
subaerially presents significant differences. The phase
Fig. 3 -
Ice avalanches due to the failure of the front of
glaciers on the oceanic water are likely to be a
special case of strongly surfing landslides. Perito
Moreno glacier, Argentina. Image Shutterstock,
reproduced with permission
Fig. 2 - Few frames from the
video by M. Pom-
pilio that show the
2002 landslide at
Stromboli volcano.
The whole sequence
lasts about 4 seconds
background image
P. MAZZANTI & F.V. DE BLASIO
428
International Conference Vajont 1963-2013. Thoughts and analyses after 50 years since the catastrophic landslide Padua, Italy - 8-10 October 2013
the following features:
i) plates thickness (the thickest plate was seldom
surfing);
ii) velocity;
iii) other tiny differences in the angle of impact,
which are poorly controllable.
THEORETICAL ESTIMATES
AIR TO WATER TRANSITION AND SURFING
PROCESS: BASIC PRINCIPLES
Several studies and experiments have been car-
ried out since
Von
K
arman
(1929) to find the impact
forces of simple shaped objects such as spheres or
disks against water. The sudden change of the ambient
medium significantly affects both the further motion
of the object and the water surface. Experiments show
the complex mechanisms related to the transition
mainly in terms of air cavitation, i.e., generation of
air bubbles behind the object plunging in water during
the initial stages of its submerged movement (P
leSSet
& P
roSPeretti
, 1977; l
ee
et alii, 1997) and shock on
the impacting body (F
aSanella
et alii, 2003). A land-
slide is much larger than any experimental object,
and also irregularly shaped. For this reason, the air-
to water impact is more complex to understand and
reproduce through laboratory experiments. Notwith-
standing the importance of the air to water transition,
little is known about the forces acting on an impacting
landslide. The impact of intact rocky blocks against
water have been considered from a dynamical view-
point, albeit in a much approximated way (D
e
b
laSio
& m
azzanti
2010; m
azzanti
& D
e
b
laSio
, 2011). In
general terms, an object colliding against water with a
velocity component normal to water v
is subject to a
vertical impact force of the form
where S
EFF
is the effective impact surface (i.e., the wet-
ted area cut horizontally through the body), ν
is the
landslide speed perpendicular to water, C
IMP
is an impact
drag coefficient, and ρ
w
is water density.
In practice, despite the relatively simple form of
equation (1), its application to the case of a landslide
impact may be more difficult than what one might ex-
pect. First of all, the impact drag coefficient is unknown
for most situations and velocities. In a previous rockfall
problem D
e
b
laSio
& m
azzanti
(2010) used the Shiff-
man-Spencer model for the drag (S
hiFFmann
& S
Pencer
,
not considered. On the other hand, preliminary experi-
ments in a small flume partially immersed in a water
tank showed a significant surfing effect of the granular
material at the air-to water transition, similar to that ob-
served in real landslides (m
azzanti
& D
e
b
laSio
, 2011).
Specifically, different inclinations of the flume and three
granular sizes were experimented. These simple tests
showed a limited influence of the impact on coarse
round grains, while evident surfing was observed with
fine round grains (m
azzanti
& D
e
b
laSio
, 2011).
A sudden reduction in velocity upon impact re-
sulted in the rising of the granular mass and a con-
sequential thickening of the frontal part of the flow,
with particles floating in the original mass. Over time,
a significant increase of the area affected by floating
was observed such that, following the impact, a por-
tion of the mass was surfing at water level.
Surfing effects by using flat grains of argillite
were even more evident. Single grains at the air-to
water interface could topple and jump, sometimes
nearly 2 m ahead of the impact point. Particles that
did not jump were capable of floating and whirling
into the water tank at high speed, sometimes travelling
through meters-long distances. However, the irregular
shape of the grains did not allow to perform controlled
analysis and to attempt a scaling investigation.
Therefore, for the present work about 50 new
tests were performed during the summer 2012 at the
geotechnical laboratory of the University of Rome
“Sapienza” (Department of Earth Sciences). The ex-
perimental setup consisted of a Plexiglas flume 1.5 m
long and 15 cm wide, partially immersed in a pool
filled by water and three disks of clay having density
1.3 kg/m
3
, radius 10.5 cm and variables thicknesses
(this leading to the following masses: 46.6 g, 90.5 g
and 117.3 g). Figure 4 shows two typical experiments
with the medium thickness disk of mass 90.5 g.
The experiment in A shows the behavior of the
disk impacting the water pool at a speed of about 3
m/s. The plate is surfing to a long distance, yet always
remaining submerged. In the experiment shown in B,
the speed has been increased to about 5 m/s. In this
case, the frontal drag force causes the plate to stand
up after impact with water, and even to topple. This
strong energy dissipation perturbs much the water sur-
face, which makes the plate stop and promptly sink.
A qualitative analysis of the experiments allowed
observing that surfing of disks is mainly controlled by
(1)
background image
LANDSLIDES SURFING ON WATER: A PRELIMINARY STUDY
Italian Journal of Engineering Geology and Environment - Book Series (6) www.ijege.uniroma1.it © 2013 Sapienza Università
Editrice
429
ficient (typically C
SK
≈ 10
-3
-10
-2
; C
IMP
≈ 1), the contribu-
tion in Eq. (3) is much smaller than the one in Eq. (2).
A different but related phenomenon can potentially
shed light on this hypothesis. The most spectacular and
dangerous mass flows which exploit a kind of surfing
process are pyroclastic flows. As already mentioned,
some papers have considered the surfing of pyroclastic
flows reaching the sea through theoretical analyses, labo-
ratory experiments (F
reunDt
, 2003; l
egroS
& D
ruitt
,
1999) and back analysis of real cases like the 1883 Kraka-
tau eruption (c
areY
et alii, 1996). Once a pyroclastic
flow reaches the sea, several mechanisms take place like
phreatic explosions mainly related to thermodynamic
phase transition from water to vapor, which may change
the flow behavior and make it flow horizontally. Cata-
strophic consequences are exemplified by the cases of
Krakatau 1883 and Vesuvius 79 A.C (c
areY
et alii, 1996).
The burning dust from Krakatau, moving over the sea,
reached some islands several km away from the volcano
maintaining high velocity and lethal temperatures. These
phenomena represent a dramatic example of the impor-
tance of water surface in controlling the evolution of natu-
ral gravity flows coming from a subaerial environment.
SURFING OF A RIGID BODY: A SIMPLE MODEL
In order to simulate the impact of a landslide onto
a water reservoir (5A) we make some theoretical esti-
mates of a simple object (such as the one shown in 5B)
falling against water perpendicular to its base with a
velocity v. The object consists of a cylinder of radius
R
i
and height T
2
with a basal spherical sector of height
T
1
and radius of curvature R.
1945; m
oghiSi
& S
quire
, 1981). However, the precise
values at high speed even for a simple sphere or cylinder
are poorly known.
Moreover, the permeability of the landslide mate-
rial is strongly affected by the presence of fines clog-
ging the pores of the fragmented landslide material, and
it is difficult to understand whether the basal layers of
a fragmenting landslide are impermeable to water. In
addition, the landslide material is deformable and as-
sumes complicated shapes during and after the impact.
For simplicity, in a first estimate we shall be con-
sidering the landslide material as rigid, at least during
the first stage of the impact. From Eq. (1), the effective
vertical acceleration g
EFF
upon impact with the water
has the functional form of the kind
where V is the volume of the model landslide, ρ
w
,
ρ are the densities of water and of the landslide re-
spectively, ΔV is the submerged volume, C
D
is an
impact drag coefficient. Eq. 2 shows that the gravity
acceleration may be strongly reduced by the impact
(second term on the right hand side) in addition to
buoyancy of the submerged part.
The strong reduction of the vertical acceleration
does not have an equivalent counterpart concerning
the horizontal acceleration, which will be of the kind
where u is the horizontal velocity (parallel to the water
level). Owing to the smaller value of the skin friction
drag coefficient C
sk
compared to the impact drag coef-
Fig. 4 - Examples of two sequences
A1-A3 and B1-B3 of sim-
ple experiments with a flat
cylindrical plate built with
modeling clay. The flume
has an inclination of 35°.
The only difference be-
tween sequence A and B
is the disk velocity at the
impact with water (3 m/s
for A and 5 m/s for B). Note
that in the “A” sequence
the plate is surfing to a
long distance, yet always
remaining submerged. In
“B”, the frontal
drag force
makes the plate to topple
(2)
(3)
background image
P. MAZZANTI & F.V. DE BLASIO
430
International Conference Vajont 1963-2013. Thoughts and analyses after 50 years since the catastrophic landslide Padua, Italy - 8-10 October 2013
This geometry is similar to disks used in the ex-
periments of Figure 4 and can be useful to analyse the
impact of more complex landslide geometries. Fur-
thermore, it is noteworthy that the explicit introduc-
tion of the radius of curvature at impact may allow
to explain the importance of the angle of impact of a
landslide in the reservoir for surfing to occur.
The total volume of this object is:
The equation of motion of the object (neglecting
the added mass effect) is
The factor Δ
v
/
v
is the ratio between the sub-
merged volume and the total volume of the object as a
function of the depth reached by the landslide.
Having approximated in this simple model the
shape of the landslide at the contact with water as the
portion of a sphere, the cross-section surface is ap-
proximated as the one of a spherical sector and so
where Γ is the depth reached by the sphere, and
We now assume that the impact with water oc-
curs with the object moving at an angle β with re-
spect to the horizontal, and not vertically. The ratio
between the impact and the gravity force (accounting
also for buoyancy) is so
Figure 6 shows the ratio between the impact and
the gravity forces as a function of the depth of the
modelled landslide speed for different radii of curva-
ture, assuming that the base of the object is submerged
by 5 m of water. Note that the ratio approaches unity
for large radii of curvature and high speeds (the impact
force increases with the square of the speed, as evident
from the equation 8). The relevant role of the radius
of curvature is also easily explained: when the radius
of curvature is small, the area of the submerged land-
slide remains small as the object dips into the water
reservoir, while for a large radius it strongly increases,
allowing for an effective increase of the impact force.
This is probably a key effect in surfing.
Coming back to the landslide, 5A and 5B show
that the same landslide falling into the water reservoir
may impact with two different effective radiuses of
curvature depending on the angle of impact: 5C shows
an impact with relatively small angle, corresponding
to large radius of curvature, while 5D an impact with
small radius of curvature. From the previous analysis,
the second landslide benefits more of the surfing effect
from the impact energy (Fig. 6).
The difference between the two situations may be
large for a landslide that is very flat but with a sharp
tip. This also indicates that the angle of impact affects
dramatically the surfing effect.
It is worth to note that a landslide may benefit from
the surfing effect even if the ratio η is less than one. This
is because a landslide that has much reduced vertical
acceleration may avoid contact with the bottom of the
sea or lake and thus hydroplane; the missing contact
with the basin will promote mobility in this case.
Fig. 5 - A: a landslide falling onto a water reservoir. B: the
approximated shape used for a simple theoretical
analysis. C: The impact at low angle corresponds
to an impact of an object like in B with large radius
of curvature. In a more head-on impact (D), the
landslide impinges with small radius of curvature
Fig. 6 - The ratio η between the impact to gravity forces
as a function of the speed for an object shaped
like that in 5B plunging at an angle β=45
0
,
C
d
=1.5 p=2200kg/m
3
T
1
=10 m; T
2
=8 m. The
data refers to the point at which the object is im-
mersed in 5 m of water. The density less than that
of basaltic rock is justified considering that the
rocks of Stromboli are often pumiceous
(4)
(5)
(6)
(7)
(8)
background image
LANDSLIDES SURFING ON WATER: A PRELIMINARY STUDY
Italian Journal of Engineering Geology and Environment - Book Series (6) www.ijege.uniroma1.it © 2013 Sapienza Università
Editrice
431
C
D
is the drag coefficient. In accordance with S
hiFFman
& S
Pencer
(1945) and m
oghiSi
& S
quire
(1981), C
D
-
varies between the situation at impact with the water
surface when it attains its highest value, and the situ-
ation of total immersion. Models of rock avalanches
based on molecular dynamics have been widely used
for the description of dry granular flows (c
amPbell
,
1990; c
amPbell
et alii, 1995). In previous work, an
attractive force was introduced to simulate a cohesive
debris flow (D
e
b
laSio
, 2009). The model presented
here also features the impact and drag with water (first
and last terms on the right hand side of Eq. 11).
By testing different plausible values of density, impact
geometry and impact coefficients, we found that the
results vary greatly between situations in which surfing
is favoured (“strong” surfing) to those of weak surfing.
In Figure 7 the case of a configuration experiences
surfing is presented. More specifically, Figure 7A shows
the simulated sequence of a landslide collapsing into
a water reservoir at different times (the water is at “0”
height). The granular mass is led to collapse against a
gate that is than “removed” numerically after 20 s. After
removal, the landslide slips along a relatively steep chute.
When reaching the water surface, blocks tend to travel
more horizontally, thus remaining at the front of the land-
slide. Then, new falling blocks form a partly submerged
plug with tendency to grow on the subaerial side, and
which is thrust horizontally by the falling blocks. Figure
7B at the bottom shows the velocity as a function of time.
Figure 8 shows the corresponding results for a weak
surfing case (obtained lowering the impact coefficient
and increasing effective gravity, i.e., gravity acceleration
acting upon a completely submerged body). Note that in
this case the landslide tends to plunge more promptly
into the water without forming a subaerial plug.
DISCUSSION AND CONCLUSIONS
We have suggested that once a landslide mass im-
pacts on the water surface, the transition between the
subaerial and the subaqueous environment could cause
a drastic reduction of the vertical acceleration. As a
consequence, the momentum of the landslide is largely
modified after the impact and transition phase and the
velocity become much more parallel to water than it
was before impact. This process, here called “surfing”,
is suggested by photographic sequences of past coastal
landslides, laboratory experiments, mathematical anal-
ysis and numerical simulations. However, not all the
Theoretically, a necessary condition for surfing
is that upon impact the total vertical acceleration be-
comes directed against the gravity field, which leads to
where v
CRIT
is the critical velocity for surfing. Hence,
for plausible parameters of a 10 m deep landslide,
critical velocities would be of the order 20 m/s.
However, a landslide such as Gongjiafang was
surfing at speeds 50% lower that the critical value
calculated with Eq. (9). Apart from uncertainties
that may justify a deviation of this order probably
we must consider that surfing occurrence is probably
easier that the eq. 9 would suggest.
Note that a vertical acceleration upon impact
does not necessarily imply a bouncing of the land-
slide, because even if the acceleration changes sign,
the velocity may not do so.
SURFING OF A DEFORMABLE BODY: A NU-
MERICAL SIMULATION
In order to improve the simple above analysis,
we have also developed a numerical model for a
landslide impacting against a water reservoir using
a molecular dynamics approach (c
amPbell
, 1990).
This method consists of substituting the debris mass
with a set of particles, each representing a block,
modelled as two-dimensional disks. The model thus
allows in a natural way for the deformation of the
landslide mass. In a previous version for subaerial
landslides, the force
t
i
(
x
i
-
x
j
)
acting on the block due
to contact with block is written as
where
x
i
and
x
j
are the positions of the two blocks.
The total force acting on i results from four elementary
units: i)
r
i
(
x
i
-
x
j
)
a repulsive force simulating the hard-
core repulsion when blocks interpenetrate, ii)
a
i
(
x
i
-
x
j
)
an attractive force simulating cohesion, iii) a shear
Coulomb force
c
i
(
x
i
-
x
j
) resulting from friction, and fi-
nally iv) a dissipative force
D
i
(
x
i
-
x
j
).
The acceleration of the block is so
where M is the block mass, ∆v its submerged volume,
F
G
is the impact force with the terrain, R is the block ra-
dius, g is gravity acceleration, and the last term is intro-
duced here to deal with water impact and flow, where
(9)
(10)
(11)
background image
P. MAZZANTI & F.V. DE BLASIO
432
International Conference Vajont 1963-2013. Thoughts and analyses after 50 years since the catastrophic landslide Padua, Italy - 8-10 October 2013
probably the most relevant features that can facilitate
the occurrence of the process:
i) low density of the moving mass;
ii) rounded snout of the mass;
iii) impermeable material due to fines clogging the
pores;
iv) high speed at the impact against water with gentle
slope angle.
We speculate that surfing of a landslide on the wa-
ter surface can significantly affect also the underwater
propagation of the landslide and, consequently the in-
duced tsunami generation.
landslides may potentially give origin to this kind of
process. For example, it is unlikely that the very thick
Vajont landslide has surfed during the event of 1963.
Different factors like the speed, impact angle,
thickness, density and permeability of the mass are
likely to strongly affect the impact with water of a
landslide starting subaerially and plunging into a
water basin. However, the controlling factors are at
present difficult to be constrained quantitatively.
We suggest that surfing will likely occur if several
favorable conditions are satisfied, some of which have
been addressed in the present work. The following are
Fig. 7 - top: sequence of computer simula-
tion of a landslide collapsing into a
water reservoir (the water is at “0”
level) in a case of favoured surfing.
The panels show the behavior at
the beginning (prior to the removal
of the numerical gate) and then at
times of 4.5 s, 11.4 s, 19 s, 25 s, 34
s, 44 s, 76 s. Bottom: the absolute
value of the average velocity as a
function of time. The impact coeffi-
cient is 2.5, the disk radius is 2 m,
and thickness (from which the mass
is calculated) is 1 m. The drag coef-
ficient 0.8; effective gravity is 3 m/s
2
background image
LANDSLIDES SURFING ON WATER: A PRELIMINARY STUDY
Italian Journal of Engineering Geology and Environment - Book Series (6) www.ijege.uniroma1.it © 2013 Sapienza Università
Editrice
433
influenced by the inclination, velocity and type of
material. Large bubbles created at the impact by
cavitation may develop with coarse round grains,
but are probably absent with smaller grains.
The surfing effect was not discussed by F
ritz
et
alii (2003) even though it seems to occur in their ex-
periments especially with fine and flat coarse grains. A
large volume of the hydrodynamic impact crater can
be observed exactly when the surfing effect does not
occur (m
azzanti
& D
e
b
laSio
, 2010). Hence, it can
be conjectured that the experiments showing hydro-
dynamic cratering as the leading effect (F
ritz
et alii,
To clarify how the water-impact can affect the
tsunami potential, two different phases could be con-
sidered separately: 1) the impact and, 2) the subse-
quent flow behavior.
Regarding the impact phase, F
ritz
et alii (2003)
showed that a thick landslide impacting against wa-
ter at high velocity creates a hydrodynamic crater
which significantly increases the volume of dis-
placed water and hence the wave height. In their ex-
periments, these authors use constant slope angle of
45° and one single type of granular material. How-
ever, the hydrodynamic impact crater is strongly
Fig. 8 - A: sequence of computer simu-
lation in a case of weak surfing.
The panels show the behavior at
times of 4.5 s, 8 s, 11 s, 15 s, 20 s,
33 s, 44 s. B: the absolute value
of the average velocity as a func-
tion of time. The impact coeffi-
cient is 2.0, the drag coefficient
0.8; effective gravity is 5 m/s
2
background image
P. MAZZANTI & F.V. DE BLASIO
434
International Conference Vajont 1963-2013. Thoughts and analyses after 50 years since the catastrophic landslide Padua, Italy - 8-10 October 2013
2003) are a consequence of the coarse round grains
used for the experiments, and cannot be generalized to
different types of landslides.
In this work, however, we concentrated more on
the subsequent flow behavior (point 2). After a land-
slide has impacted against the water it may either
promptly descend into the reservoir, or surf (complete-
ly or, more probably, partially upon the water surface)
before plunging into the water. We suggest that tsuna-
mis generated by surfing landslides may be different
from those due to landslides descending at once, as the
former tend to "shovel" water in the front, rather than
perturbing it from below. Thus, waves generated by a
surfing landslide will resemble more those by a fast-
moving ship, with a characteristically different dy-
namics. It would be interesting to understand whether
the surfing effect, which can significantly reduce the
hydrodynamic impact crater, will consequently lessen
the tsunami potential as well, or perhaps increases it,
owing to the water horizontal movement.
The front thickening due to the impact might
also accentuate the tsunami potential (h
arbitz
et
alii, 2006; F
ritz
et alii, 2003; 2004). In addition, the
back-tilting effect of the mass might also play a non-
negligible role facilitating the injection of a water
wedge beneath the landslide body, and thus triggering
the hydroplaning effect (so delaying the contact of the
landslide with the bottom). The occurrence of hydro-
planing, with its characteristically reduced friction at
the base, significantly increases the landslide velocity
and travel distance and thus its tsunamigenic poten-
tial. It is clear that, as also shown by our simulations
(compare Figs. 7 and 8), a landslide will in any case
increase its run-out by surfing.
Perhaps surfing is a more common occurrence
than previously thought. It is possible that the scar-
city of filmed sequences of failure along the coasts
have prevented researchers from recognizing the oc-
currence and perhaps even commonness of the effect.
However, the commercial success of mobile smart-
phones and non-professional video cameras implies
an increasing chance of collecting videos of sudden
and unpredictable catastrophic events, which will
likely improve the data available.
REFERENCES
b
olin
h., Y
in
Y., l
iu
g., W
ang
S., c
hen
x. & h
uo
z. (2012) - Analysis of waves generated by Gongjiafang landslide in Wu Gorge,
three Gorges reservoir, on November 23, 2008. Landslides, 9 (3): 395-404.
b
olin
h., l
iDe
c., x
uanming
P., g
uanning
l., x
iaoting
c., h
aogang
D. & t
ianci
l. (2010) - Assessment of the risk of rockfalls
in Wu Gorge, Three Gorges, China. Landslides, 7 (1): 1-11.
c
amPbell
C.S. (1990) - Rapid granular flows. Annual Review of Fluid Mechanics, 22: 57-92.
c
amPbell
c.S., c
learY
P.W. & h
oPKinS
m. (1995) - Large scale landslide simulations: Global deformations, velocities and basal
friction. J. Geoph. Res.. 100: 8267-8283.
c
hiocci
F.l., r
omagnoli
c., t
ommaSi
P., & b
oSman
a. (2008) - The Stromboli 2002 tsunamigenic submarine slide: characteristics
and possible failure mechanisms. Journal of Geophysical Research, 113: B10102, doi:1029/2007JB005172.
D
e
b
laSio
F.V. (2011) - Introduction to the physics of landslides. Springer Verlag, Berlin.
D
e
b
laSio
F.V. (2009) - Preliminary discrete model in a computer simulation of cohesive debris flows. Geotechnical Geological
Engineering, 30: 269-276.
D
e
b
laSio
F.V. & m
azzanti
P. (2010) - Subaerial and subaqueous dynamics of coastal rockfalls. Geomorphology, 115: 188-193.
D
i
r
iSio
m., D
e
g
irolamo
P., b
ellotti
g., P
anizzo
a., a
riS
-t
oDemo
F., m
olFetta
m.g. & P
etrillo
a.F. (2009) - Landslide-
generated tsunamis runup at the coast of a conical island: new physicalmodel experiments, J. Geophys. Res., 114: C01009,
doi:10.1029/2008JC004858, 2009a.
F
aSanella
e.l., j
acKSon
K.e, S
ParKS
. c.e. & S
areen
a.K. (2003) - Water impact test and simulation of a composite energy
absorbing fuselage section. American Helicopter Society, 59
th
Annual Forum, Phoenix, AZ, May 6-8, 2003.
F
ine
i.V., r
abinoVich
a.b., b
ornholD
b.D. t
homSon
r.e. & K
uliKoV
E.A. (2005) - The Grand Banks landslide-generated
tsunami of November 18, 1929: preliminary analysis and numerical modeling. Marine Geology, 215: 45-57.
F
reunDt
a. (2003) - Entrance of hot pyroclastic flows into the sea: Experimental observations. Bull. Volcanology, 65:
144-164.
F
reunDt
a., S
trauch
W., K
utterolF
S. & S
chmincKe
H-U. (2007) - Volcanogenic tsunamis in lakes: examples from Nicaragua
and general implications. Pure and Applied Geophysics, 164: 527-545.
background image
LANDSLIDES SURFING ON WATER: A PRELIMINARY STUDY
Italian Journal of Engineering Geology and Environment - Book Series (6) www.ijege.uniroma1.it © 2013 Sapienza Università
Editrice
435
F
ritz
h.m., h
ager
, W.h. & m
inor
h.-e. (2003a) - Landslide generated impulse waves. Part 1: instantaneous flow fields, Exp.
Fluids, 35: 505-519. doi:10.1007/s00348-003-0659-0.
F
ritz
h.m., h
ager
W.h. & m
inor
h.-e. (2003b) - Landslide generated impulse waves. Part 2: hydrodynamic impact craters.
Exp. Fluids, 35: 520-532.
K
remer
K., S
imPSon
g. & g
irarDcloS
S. (2012) - Giant Lake Geneva tsunami in AD 563. Nature Geoscience. DOI: 10.1038/
ngeo1618
l
ee
m., l
ongoria
r.g. & W
ilSon
D.e. (1997) - Cavity dynamics in high-speed water entry. Phys. Fluids, 9: 3.
m
azzanti
P. (2008) - Analysis and modelling of coastal landslides and induced tsunamis. Ph.D. thesis, “Sapienza” University of
Rome, Department of Earth Sciences, 212 pp.
m
azzanti
P. & D
e
b
laSio
F.V. (2011) - The dynamics of coastal landslides: insights from laboratory experiments and theoretical
analyses. Bull Eng Geol Environ, 70: 411-422, DOI 10.1007/s10064-010-0322-y.
m
azzanti
P. & b
ozzano
F. (2011) - Revisiting the February 6
th
1783 Scilla (Calabria, Italy) landslide and tsunami by numerical
simulation. Marine Geophysical Research, 32: 273-286, DOI 10.1007/s11001-011- 9117-1, ISSN: 0025-3235.
m
oghiSi
m. & S
quire
, P.t. (1981) - An experimental investigation of the initial force of impact on a sphere striking a liquid
surface. J. Fluid Mech., 108: 133-146.
m
ohammeD
F. & F
ritz
h.m. (2010) - Experiments on Tsunamis Generated by 3D Granular Landslides. Submarine mass
movements and their consequences, 28: 705-718.
P
leSSet
m.S. & P
roSPeretti
a. (1977) - Bubble dynamics and cavitation. Ann. Rev. Fluid Mech., 9: 145-185.
S
chuSter
r.l. (1979) - Reservoir-inducedland-slides: Bulletin of the International Association of Engineering Geology, 20: 8-15.
S
emenza
E. (2002) - La storia del Vajont, raccontata dal geologo che ha scoperto la frana. Tecomproject, Ferrara.
S
hiFFman
m. & S
Pencer
, D.c. (1945) - The force of impact on a sphere striking a water surface. Approximation by the flow about
a lens. Applied Math. Panel, National Defense Research Committee, Report No. 105.
t
inti
S., m
aramai
a. & F
aVali
P. (1995) - The Gargano promontory: an important Italian seismogenic-tsunami-genic area. Mar.
Geol., 122: 227-241.
background image
Statistics