Document Actions

ijege-13_bs-vassallo-et-alii.pdf

background image
Italian Journal of Engineering Geology and Environment - Book Series (6) www.ijege.uniroma1.it © 2013 Sapienza Università
Editrice
371
DOI: 10.4408/IJEGE.2013-06.B-35
MONITORING OF THE MOVEMENTS OF
A DEEP, SLOW, CLAYEY LANDSLIDE AND 3D INTERPRETATION
R
obeRto
VASSALLO
(*)
, R
ossella
PAGLIUCA
(*)
& C
ateRina
DI MAIO
(*)
(*)
University of Basilicata - School of Engineering - Via dell’Ateneo Lucano 10 - 85100 Potenza, Italy
E-mail: roberto.vassallo@unibas.it - Tel. +390971205390
The landslide is from very to extremely slow. In fact,
in the monitoring period, the displacement rates have
always been in the order of one cm/year in the middle
of the track, where only a few houses have been built
in the last twenty years, and seem still unaffected by
the slope movements. On the contrary, at the land-
slide head - which is the fastest zone - a house has
recently been evacuated because it is threatened by
the landslide and by a shallow earth-flow occurring
in the main scarp. The landslide foot, which is the
slowest part, with average rates in the order of a few
mm/year, is crossed by a highway and by the national
railway and several buildings arise on it.
With the exception of the head of the landslide,
the rates of displacement are too low to vacate inhab-
itants or deviate roads and railway. In addition, the
landslide is very deep and thus very difficult to sta-
bilize. On the other hand, a deep understanding of the
landslide, which permits to predict its behaviour, can
contribute to minimize the associated risk.
In order to understand how and when the land-
slide movement can cause damage to houses and in-
frastructures, inclinometer tubes for both standard,
periodical measurements and continuous data acquisi-
tion were installed since 2004 (Fig.1). GPS permanent
and periodical stations were also installed. Piezometer
measurements, in situ permeability tests and labora-
tory mechanical tests were carried out.
The maximum detected depth of the slip sur-
face resulted of about 40 m. Displacement profiles
ABSTRACT
The Costa della Gaveta landslide is an active, deep
and very slow landslide which is being monitored and
studied since 2004. Recently new boreholes have been
carried out and inclinometer tubes have been installed
to deepen the comprehension of the 3D geometry and
kinematics. Localized displacement on a slip surface is
confirmed to be the prevailing movement mechanism.
Thanks to the long term monitoring, even the slow in-
ternal viscous deformations of the landslide body have
been evaluated and interpreted. Soil discharge (i.e., the
volume rate of soil crossing transversal sections) is con-
firmed to be constant in the track. This characterizing
kinematic feature allows a first approximation predic-
tion of the landslide behaviour on empirical basis.
To further improve the prediction of the landslide
behaviour and the risk management of the area, a
theoretical model has been constructed by a 3D finite
element code. First results show the large difference in
the safety factor between 2D and 3D analyses for the
considered landslide. The influence of pore pressure
distribution on the local safety factor is also analysed.
K
ey
words
: landslide, displacements, inclinometer, predic-
tion, stability
INTRODUCTION
The Costa della Gaveta landslide occurs close to
the city of Potenza, Italy, in a structurally complex
clay formation, locally known as Varicoloured Clays.
background image
R. VASSALLO, R. PAGLIUCA & C. DI MAIO
372
International Conference Vajont 1963-2013. Thoughts and analyses after 50 years since the catastrophic landslide Padua, Italy - 8-10 October 2013
mented with the aim of evaluating the stress-strain
distribution within the landslide body and its varia-
tions induced by the possible triggering factors. The
difference in the safety factor provided by a 2D and a
3D model is also analysed.
SUBSOIL CHARACTERIZATION
The landslide occurs within the geological for-
mation of the Varicoloured Clays (Upper Cretaceous
- Oligocene), which is made up of a succession of
chaotic, severely tectonized scaly clays and marly
clays. The formation includes rock fragments or
blocks or even strata of marly limestone, calcareous
marl and calcarenite. The soil is markedly heteroge-
neous, similarly in the landslide body and in the sta-
ble soil underneath (D
i
M
aio
et alii, 2010).
Tab. 1 reports the ranges of some physical char-
acteristics of the finer, clayey matrix. Because of the
diffuse presence of rock fragments, only a few undis-
turbed samples could be extracted from the boreholes
indicated in Fig. 1. On these samples, the peak shear
strength of the landslide material was evaluated by
isotropically consolidated - undrained (CIU) triaxial
tests. The residual strength was evaluated by direct
shear tests on both the undisturbed and the reconsti-
tuted material. The details of the experimentation are
reported in D
i
M
aio
et alii (2010). The peak strength
is characterized by average c’ = 50 kPa and φ′ = 14°,
while the residual strength is characterized by c’ = 0
and average φ′
r
of about 10°, as reported in Tab. 1.
Thus, a noticeable drop of strength from peak to re-
sidual is observed. In the finite element analyses, the
available strength on the slip surface will be hypoth-
obtained in the first years - mostly along the land-
slide median longitudinal section - were practically
uniform in each inclinometer from the slip surface to
the ground: sliding along the slip surface was thus
considered the prevailing mechanism of displace-
ment. Furthermore, in the observation period, the
displacement rates were always noticeably decreas-
ing from upslope to downslope. Under the hypothesis
that displacements were uniform not only along the
inclinometer vertical but in each entire transversal
section of the track, D
i
M
aio
et alii (2010) interpreted
the decrease of displacement rate as an effect of the
increase in the areas of transversal sections in the
downslope direction, the resulting “soil discharge”
(i.e., the volume of soil crossing transversal sections
in the unit time) being practically constant.
In the first half of 2012 new boreholes have been
carried out and inclinometer tubes have been installed
to study the kinematics of a transversal section of
the track. Based on previous and new data, this pa-
per proposes a 3D empirical model of the landslide.
The contribution of internal viscous deformations
to ground displacements and soil discharge are also
considered. Furthermore, a 3D finite element (FEM)
theoretical model of the landslide has been imple-
Fig. 1 - Landslide boundaries, inclinometers, piezom-
eters, GPS stations, and considered longitudinal
and transversal sections
Tab. 1 - Physical characteristics, index, hydraulic and me-
chanical properties of the finer part of the land-
slide material
background image
MONITORING OF THE MOVEMENTS OF A DEEP, SLOW, CLAYEY LANDSLIDE AND 3D INTERPRETATION
Italian Journal of Engineering Geology and Environment - Book Series (6) www.ijege.uniroma1.it © 2013 Sapienza Università
Editrice
373
2006, when the slip surface in it had been clearly de-
tected, three fixed-in-place probes were installed, one
in correspondence to the slip surface and the other two
in the landslide body and in the stable soil respectively
(Fig. 2). In January 2009, the fixed-in-place probes
went out of use. In July 2010, they were removed and
inclinometer measurements were carried out by mo-
bile probe again. A very good agreement was found
between displacements measured in such different
ways, as it is possible to appreciate from data of I9
reported in Fig. 3. At the beginning of 2012, the tube
was found pinched off in correspondence to the slip
surface, so, that of December 2011 is the last measure-
ment carried out beneath the slip surface. Inclinom-
eters I11, I8 and I7 are still in full use today.
Figure 2 shows that three components of displace-
ment can be clearly distinguished and analysed in
each inclinometer profile: AB, corresponding to slid-
esized to be close to the residual, consistently with
the observation that this active and ancient landslide
underwent large displacements along a regular slip
surface. Hence, if this hypothesis is correct, the shear
strength within the landslide body is much higher than
that on the slip surface. As reported by D
i
M
aio
et alii
(2010) and V
assallo
et alii (2012a-b) such feature is
recognized in the literature as an important factor de-
termining the style of displacements and justifying the
prevailing contribution of sliding at the base.
Hydraulic conductivity was estimated in situ by
carrying out falling head tests, and in laboratory by in-
terpreting the consolidation curves of oedometer tests.
The estimated conductivity values, reported in Tab. 1,
are very low and this strongly influences the landslide
response to rainfall (V
allaRio
, 2012).
INCLINOMETER MEASUREMENTS
Five inclinometer casings, up to 50 m deep, were
installed in the landslide at the end of 2004, in the po-
sitions indicated by Fig. 1. In 2006, a GPS network of
permanent and non-permanent stations was installed
for the evaluation of surface displacements. In the
observation period, a good consistency was found
between inclinometer and GPS measurements (C
al
-
CateRRa
et alii, 2012). Other inclinometers were in-
stalled recently, in the first half of 2012.
A selection of the many measurements carried out
from 2005 to today is reported in Fig. 2 in terms of in-
clinometer cumulative displacement profiles. Casing
I10 was found out of use in August 2008. Casing I9
was used for different kinds of measurements. In July
Fig. 2 - Cumulative displacement profiles of casings installed close to the axial longitudinal section
Fig. 3 - Deep displacements against time
background image
R. VASSALLO, R. PAGLIUCA & C. DI MAIO
374
International Conference Vajont 1963-2013. Thoughts and analyses after 50 years since the catastrophic landslide Padua, Italy - 8-10 October 2013
mums differing of less than an order of magnitude.
So, it seems possible to refer to an average almost
constant displacement rate which, in the case of I9, is
about 1 cm/year.
As for the trends of the BC and CD components,
reported in Fig. 5, one can observe their low values
and an average continuous increase in time. D
i
M
aio
et alii (2013) propose an interpretation of their general
trends based on Bingham’s rheological model and on
the results of long term laboratory shear creep tests.
At the beginning of 2012 two new inclinometer
tubes, I9b and I9c, have been installed in the transver-
sal section through I9, in the position indicated in Fig.
1. Because of the extremely slow rate of movements,
at the moment, these two inclinometers provide only
qualitative information susceptible to variations. How-
ever, first results seem consistent with the measure-
ments of I9 in terms of slip surface position, as will be
shown in the next section. Another interesting feature
is that in these lateral boreholes internal deformations
seem more significant than those of the median lon-
gitudinal section and contribute to a noticeable per-
centage of total displacements. D
i
M
aio
et alii (2013)
report the water content (w) profiles of I9b and I9c
showing that in both boreholes the slip surface is close
ing along the slip surface, BC and CD, resultants of
internal deformations of the landslide body, which are
more pronounced in the upper 2-2.5 m thick soil layer,
consisting of remoulded organic soil characterized by
poor mechanical properties (D
i
M
aio
et alii, 2013).
The sliding component AB largely prevails on
the others, also in the inclinometer I12 recently in-
stalled, and thus gives the character of substantial
uniformity to the profiles.
Figure 3, which reports deep displacements on the
slip surface against time, shows that the AB compo-
nents are strongly correlated to each other in the whole
observation period. In particular, it can be observed
that displacements in I9 are very close to those of I8
multiplied by 2. Furthermore, displacements of I8 mul-
tiplied by 4 are in good agreement with both those of
I10 and those of I12. It can be said that, with the excep-
tion of the period of apparent acceleration that led I10
to failure, deep displacements of any two inclinometer
casings keep an almost constant ratio (D
i
M
aio
et alii,
2013). This is actually one of the first considerations
on the basis of which D
i
M
aio
et alii (2010) hypoth-
esized the constancy of the soil discharge.
The displacement rates, which correspond to the
slope of the curves in Fig. 3, range from 2 - 3 cm/year
in the upper portion of the track to about 1 mm/year
in the accumulation. Actually, the extremely slow dis-
placements measured in I7 have been considered re-
alistic because of their regular evolution over 7 years,
besides their clear azimuth trend.
It is worth noting that in the period 2006-2009 the
landslide rarely came to a stop. This is shown by Fig.
4, where the displacement rate in I9 in that period is
computed from the measurements of the central fixed-
in-place probe, installed in correspondence to the slip
surface. The rate went rarely down to zero, and always
attained very small values, with minimums and maxi-
Fig. 4 - Displacement rate on the shear surface in I9
and I8 against time, obtained from fixed-in-place
probe and mobile probe measurements
Fig. 5 - Displacements against time: BC and CD components (defined in Fig. 3)
background image
MONITORING OF THE MOVEMENTS OF A DEEP, SLOW, CLAYEY LANDSLIDE AND 3D INTERPRETATION
Italian Journal of Engineering Geology and Environment - Book Series (6) www.ijege.uniroma1.it © 2013 Sapienza Università
Editrice
375
head. Although the details of the landslide geometry
in this zone are still under study, the depth of failure
has been clearly detected (Fig. 7a).
Measurements in lateral inclinometers I9b and I9c
contribute to investigate the geometry of a transversal
section, as shown by Fig. 8, where they are compared
to the data of the central inclinometer I9. The meas-
urements in I9c indicate a position of the slip surface
very close to that hypothesized by D
i
M
aio
et alii
(2010), whereas, in I9b, the slip surface is detected 3
m above the hypothesized depth. So, the geometrical
reconstruction proposed by D
i
M
aio
et alii (2010) for
the three cross sections through the inclinometers I8,
I9 and I10 can be reasonably extended to the whole
track. Figure 9 reports a simplified scheme of the slip
surface both in the landslide track and in the accumu-
lation, which takes into account experimental data,
theoretical elaborations and simplification require-
ments of numerical calculation. Two planes, a and b,
were used to describe the slip surface in the track. They
to the boundary between less consistent (w
=
25%) and
more consistent (w
=
15%) soil, as shown in Fig. 6,
where displacement profiles and water content data are
compared. On the contrary, in I9 the slip surface goes
through the more consistent material. The different
water content is thus considered a possible explanation
of the different viscous deformations.
GEOMETRICAL FEATURES
The landslide is characterized by a large well-de-
fined fan-shaped foot, a track with regular flanks and
a wide depletion zone (Fig. 1). The main geometrical
features of the landslide were described by D
i
M
aio
et
alii (2010). In particular, the trace of the slip surface in
the median longitudinal section was drawn on the basis
of the inclinometer profiles (Fig. 7a). The landslide de-
tected depth increases from about 10 m in inclinometer
I10 to 38 m in inclinometer I8. This latter can be con-
sidered at the transition from the track to the accumula-
tion, which extends down to the Basento river and is
partly subjected to its erosion (V
assallo
et alii, 2012a).
The geometry of some cross sections of the landslide
body (Fig. 7b) were drawn by D
i
M
aio
et alii (2010)
on the basis of the inclinometer profiles and of the
elaboration of geometrical characteristics of the flanks
of the track, which were considered as the outcropping
portion of the slip surface. The resulting cross section
areas were found strongly increasing in the downslope
direction. This has been considered as a cause of the
significant decrease in the average displacement rate
from I10 to I7 clearly shown by the data in Fig. 3.
The inclinometers recently installed add further
information about the landslide geometry and kin-
ematics. Inclinometer I12 is located at the landslide
Fig. 6 - Cumulative displacement profiles and water con-
tent profiles of I9b and I9c
Fig. 7 - Slip surface in the longitudinal median section (a)
and in some transversal sections (b)
Fig. 8 - Displacement profiles of inclinometers I9, I9b and
I9c and comparison with the slip surface hypoth-
esized by D
i
M
aio
et alii (2010)
background image
R. VASSALLO, R. PAGLIUCA & C. DI MAIO
376
International Conference Vajont 1963-2013. Thoughts and analyses after 50 years since the catastrophic landslide Padua, Italy - 8-10 October 2013
are the best fitting planes through the lateral channels
and through the slip surface detected by inclinome-
ters. Plane c smooths off the angle formed by the first
two in order to avoid numerical problems in the FEM
model. Plane d, with an inclination greater than that
of the hill slope, describes the transition to the river
floodplain. The existence of such plane emerges from
the analysis of the ground surface topography. Its geo-
logical origin is currently under study. Plane e is the
best fitting plane of the river bank. Such scheme was
used for constructing the slip surface in the 3D finite
element model. Fig. 9 also shows the simplified slip
surface inserted into the DTM of the slope.
3D INTERPRETATION OF EXPERIMEN-
TAL DATA
The reconstruction of the 3D geometry of the
landslide and the description of its fundamental kin-
ematic features can be the basis of a prediction model.
In fact, under the hypothesis of: a) constant soil dis-
charge through the different transversal sections of
the track, and b) soil discharge independent of time,
predictions are possible, in the absence of exceptional
events such as earthquakes or extreme rainfall. In fact,
average displacement rates and then displacements
over a generic time interval can be calculated.
As for the assumption of constant soil discharge
through different transversal sections, D
i
M
aio
et alii
(2013) evaluate the soil volume crossing the transver-
sal sections against time including the creep effect. On
the basis of measured inclinometer profiles I9, I9b and
I9c and of the relative position of these boreholes in
the cross section, they hypothesized uniform displace-
ments at each given elevation for any cross section of
the track (Fig. 10). Just for the upper 2 m thick weath-
ered soil, where the shear creep rate is higher, an ad-
ditional contribution, uniform at each given depth from
the ground surface, was considered. The soil volumes
relative to the transversal section through I8 are slight-
ly higher than those relative to the sections through
I9 and I10 (Fig. 10). The volumes almost coincide if
the cross section through I8 is multiplied by 0.75. So,
assuming a constant soil discharge is confirmed to be
more than reasonable.
It is worth noting that, thus far, the volumes calcu-
lated including the creep effect are just slightly higher
than those which would be estimated by considering
the displacements uniform in each transversal section.
Nevertheless, under unchanged conditions, such dif-
ference is going to increase as years go by.
In the case of the cross section through I7, the
displacements are still very small. In I7 it is not pos-
sible to distinguish the internal deformations, with ex-
ception of those of the upper 2.5 m thick superficial
layer, and, a fortiori, it is not possible to hypothesize
a distribution of internal deformations in the section.
The sliding component does not seem uniform in the
section, as reported in detail by D
i
M
aio
et alii (2010).
The volumes through I7 reported in Fig. 10 were thus
calculated in two different ways, as schematized in
Fig. 10 and reported by D
i
M
aio
et alii (2013).
From Fig. 10, an average soil discharge in the
Fig. 9 - Three-dimensional scheme of the slip surface, consisting of: two planes (a and b) in the track area; a third plane (c)
used to smooth off the angle formed by the first two; a transition plane (d); the floodplain (e)
background image
MONITORING OF THE MOVEMENTS OF A DEEP, SLOW, CLAYEY LANDSLIDE AND 3D INTERPRETATION
Italian Journal of Engineering Geology and Environment - Book Series (6) www.ijege.uniroma1.it © 2013 Sapienza Università
Editrice
377
iour with zero cohesion and by various friction angle
values for parametric calculations. The code allows
to define the pore pressure distribution by imposing
a phreatic surface; the piezometric head is consid-
ered constant along each vertical. In the following
example, three different positions of the phreatic sur-
face are considered, one coincident with the ground
surface and the others at 2 and 4 meters below the
ground surface, respectively.
The stress distribution on the slip surface is one of
the possible outputs of the code. For any considered
case, the safety factor SF was calculated following
s
tianson
et alii (2011):
track can be evaluated: Q ≈ 0.1 m
3
/day. If our inter-
pretation is correct, then it is possible to use data ob-
tained in a few instrumented zones to estimate the dis-
placements in any other zone of the landslide. This is
shown by Fig. 11 which reports the transversal section
areas A along the longitudinal axis and the average
displacement rate (v
=
Q/A) in the transversal sections.
Actually, sliding largely prevails on creep internal de-
formations, so the real rates in any point of each trans-
versal section are very close to the average values
reported in the figure. Under the hypothesis that such
rates keep constant over a certain number of years, the
displacements can be calculated and thus the expected
damage can be evaluated (M
ansouR
et alii, 2011).
3D FEM MODEL
A 3D finite element model has been implement-
ed by the Plaxis3D code. A mesh including about
112,000 10-noded tetrahedral elements has been
adopted (Fig. 12a). The ground surface is described
by a grid with a maximum thickness of 10 m. The
landslide body is saturated even above the water ta-
ble (with unit weight γ
sat
= 21 kN/m
3
), homogeneous,
isotropic and characterized by an elastic-perfectly
plastic behaviour with a Young modulus E’ = 50 MPa
and a Poisson’s ratio ν’ = 0.40. The cohesion and fric-
tion angle of the soil are respectively c’=
50 kPa and
φ’=
14°, as from laboratory tests. The slip surface
reported in Fig. 9 has been simulated with interface
elements characterised by a perfectly plastic behav-
Fig. 10 - Volume of soil crossing the transversal sections in the monitoring period
Fig. 11 - Estimate of average displacement rate in the
transversal sections. Data relative to I12 are still
preliminary
background image
R. VASSALLO, R. PAGLIUCA & C. DI MAIO
378
International Conference Vajont 1963-2013. Thoughts and analyses after 50 years since the catastrophic landslide Padua, Italy - 8-10 October 2013
of the accumulation are at failure, independently from
each other. This seems in agreement with the physical
observation of a local instability caused by river ero-
sion at the toe (V
assallo
et alii, 2012a). Figure 12c
reports the results obtained by fixing the friction angle
value to 14° and considering three different depths of
the phreatic surface. This figure shows how the safety
factor would evolve in response to seasonal changes
in boundary conditions, under the hypothesis of an in-
stantaneous response at any depth. Actually, due to the
low hydraulic conductivity, the phenomenon is strong-
ly time-dependent. V
allaRio
(2012) carried out pore
pressure calculations in several longitudinal sections
where the sums are extended to all the triangular el-
ements of area A
i
constituting the slip surface while
τ
max,i
and τ
i
are the available shear strength and the act-
ing shear stress respectively.
Figure 12 reports, for several considered cases, the
distribution of plastic points on the slip surface. In par-
ticular, Fig. 12b reports the results obtained by fixing
the phreatic surface at the ground level and consider-
ing three different friction angle values on the slip sur-
face. It is interesting to observe that even for φ’=15°
the material in the track and in the lowermost portion
Fig. 12 - Plaxis 3D results. a) 3D mesh; b) distribution of plastic points obtained by fixing the phreatic surface at the ground
level and considering three different friction angle values on the slip surface; c) distribution of plastic points for φ'
r
=
14° and for three different positions of the phreatic surface; d) safety factor against friction angle (the 2D calculation
is relative to a phreatic surface at the ground level, 3D calculations refer to a phreatic surface with depth zw of 4m,
2m and 0 from the ground surface)
background image
MONITORING OF THE MOVEMENTS OF A DEEP, SLOW, CLAYEY LANDSLIDE AND 3D INTERPRETATION
Italian Journal of Engineering Geology and Environment - Book Series (6) www.ijege.uniroma1.it © 2013 Sapienza Università
Editrice
379
of the considered landslide, from the median axis to the
lateral channel. The Author found that in the median
section, where the slip surface reaches a depth of 38 m,
the historical rainfall series does not cause significant
pore pressure changes. Closer to the lateral channels,
where the slip surface is less deep, rain effects become
more pronounced and the safety factor exhibits appre-
ciable seasonal variations. The results reported in Fig.
12c are thus only a first step of the analysis that allows
to verify if the model works correctly. The study of the
time-dependent phenomenon is one of the real goals
of our research.
The safety factor computed for the above cases
and for other similar cases are reported in Fig. 12d.
For comparison, the safety factor evaluated by a 2D
analysis is also shown. This latter safety factor was
obtained by limit equilibrium analyses carried out by
the Slope/W code of the Geostudio suite on the longitu-
dinal medium section, with phreatic surface coincident
with the ground surface. Under equal conditions, in the
friction angle range in which the comparison is possible
(i.e., for SF>1) the global safety factors calculated by
the 3D analysis are higher than those of the 2D analysis.
Of course, the advantage of a 3D analysis is the pos-
sibility of studying the local conditions of stability, if
the model is verified to be reliable. Such verification is
currently being carried out by comparison of theoreti-
cal and experimental stress distributions.
REFERENCES
C
alCateRRa
s., C
esi
C., D
i
M
aio
C., G
aMbino
P., M
eRli
K., V
allaRio
M. & V
assallo
R. (2012) - Surface displacements of two
landslides evaluated by GPS and inclinometer systems: a case study in Southern Apennines, Italy. Natural Hazards, 61: 257-266.
D
i
M
aio
C., V
assallo
R. & V
allaRio
M. (2013) - Plastic and viscous shear displacements of a deep and very slow landslide in
stiff clay formation. Engineering Geology, 162: 53-66.
D
i
M
aio
C., V
assallo
R., V
allaRio
M., P
asCale
s. & s
Dao
F. (2010) - Structure and kinematics of a landslide in a complex
clayey formation of the Italian Southern Apennines. Engineering Geology, 116: 311-322.
F
ell
R., C
oRoMinas
J., b
onnaRD
C., C
asCini
l., l
eRoi
e. & s
aVane
W.Z. (2008) - Guidelines for landslide susceptibility, hazard
and risk zoning for land-use planning. Engineering Geology, 102: 99-111.
M
ansouR
M.F., M
oRGensteRn
n.R. & M
aRtin
C.D. (2011) - Expected damage from displacement of slow-moving slides.
Landslides, 8 (1): 117-131.
s
tianson
J.R., F
ReDlunD
D.G & C
han
D. (2012) - Three-dimensional slope stability based on stresses from a stress-deformation
analysis. Canadian Geotechnical Journal, 48: 891-904.
V
allaRio
M. (2012) - Kinematics of two landslides in Varicoloured Clays and analysis of possible triggering factors. Ph.D.
Thesis (in Italian), University of Basilicata, Potenza, Italy.
V
assallo
R., D
i
M
aio
C., C
oMeGna
l. & P
iCaRelli
l. (2012a) - Some considerations on the mechanics of a large earthslide in
stiff clays. Proc. XI Int. Symp. on Landslides, Banff, Canada, 1: 963-968.
V
assallo
R., D
i
M
aio
C., V
allaRio
M. (2012b) - A possible mechanism of movement of an ancient clayey landslide. Proc. II
World Landslide Forum, Rome, Italy, 2: 273-279.
CONCLUDING REMARKS
On the basis of a long term monitoring and of a
recent integrative investigation, the geometry and the
kinematics of the Costa della Gaveta landslide have
been reconstructed. In the observation period, the
displacement rate variations have been negligible.
So, under the hypothesis of constant soil discharge
through the different transversal sections of the track,
and of soil discharge independent of time, displace-
ment predictions on pure empirical basis are possible,
obviously in the absence of exceptional events such as
earthquakes or extreme rainfall.
A further improvement in the comprehension and
prediction of the landslide behaviour can be achieved
through a robust theoretical model. In the case un-
der consideration, a conceptually satisfying model
requires considering the 3D conditions. In fact, first
results obtained by Plaxis3D show the large difference
in the safety factor between 2D and 3D analyses un-
der hydraulic steady state conditions. Such difference
reasonably increases when the transient processes in-
duced by the historical rainfall series are considered.
The study of these transient processes is one of the
next goals of our study.
ACKNOWLEDGEMENTS
The Authors would like to thank Mr M. Belvedere
who carried out the in situ measurements from 2011 on.
background image
Statistics