Engineering Geology 196 (2015) 71–87

Contents lists available at ScienceDirect

Engineering Geology

journal homepage: www.elsevier.com/locate/enggeo

Tunnelling-induced landslides: The Val di Sambro tunnel case study

Annalisa Bandini, Paolo Berry, Daniela Boldini ⁎
Department of Civil, Chemical, Environmental and Materials Engineering (DICAM), University of Bologna, via Terracini 28, 40131, Bologna, Italy

a r t i c l e i n f o a b s t r a c t

Article history: Tunnel excavation in mountainous or sloping regions can sometimes reactivate or accelerate landslide phenom-
Received 20 March 2015 ena with a potential important impact on the stability of such constructions. The problem of tunnelling-induced
Received in revised form 30 June 2015 landslides, that represents an important issue for many areas of the world, is very complex to analyse and de-
Accepted 1 July 2015
manding to solve. In addition, there are very few well documented case histories in literature. The paper presents
Available online 6 July 2015
the case of the Val di Sambro twin tunnels excavated in a structurally complex formation, in a geomorphological
Keywords:
setting characterised by landslides of various areal sizes, whose depth was not well known before tunnelling. After
Landslide a description of the main technical features of the tunnels, the geological setting and the geomechanical site
Twin tunnels conditions, the paper illustrates and comments the most significant outcomes of the monitoring system in the
Structurally complex formations Southern zone of the tunnels, one of the areas where the major deformative phenomena occurred. The Val di
Geotechnical monitoring Sambro tunnels represent a unique case study for the large amount of monitoring data collected before and during
SAR interferometry the tunnel excavation through an integrated system, consisting of various measure methods (inclinometers,
piezometers, underground and surface topographic measurements, SAR interferometric technique). The com-
bined use of surface, subsurface, underground and satellite data allowed analysing the characteristics of the land-
slide movements before the tunnel excavation and in relation to the tunnel construction.
© 2015 Elsevier B.V. All rights reserved.

1. Introduction including roto-translational and earth-flow movements and involving

The excavation of a tunnel causes a stress redistribution inducing
deformations in the ground surrounding the opening. In areas prone to
instability, such as in the case described in this paper, the induced stress
state can trigger or accelerate landslide movements, even on large scale
(Sciotti and Calabresi, 2000; Picarelli et al., 2002; Barla et al., 2006;
Noferini et al., 2007; Urcioli and Picarelli, 2008; Wang, 2010; Bozzano
et al., 2011; Jiao et al., 2013; Sciotti and Pigorini, 2013; Maiorano et al.,
2014). In particular, Urcioli and Picarelli (2008) reviewed some Italian
case studies of the interaction between landslides and man-made
works. Wang (2010) showed a number of damage patterns of the tunnel
lining in relation to different slope instabilities. Noferini et al. (2007) and
Bozzano et al. (2011) focused on GB-SAR technique applied to monitor
landslides affected by tunnelling. Sciotti and Pigorini (2013) reported
the case study of the railway tunnel Scianina–Tracoccia that induced a
slope collapse 156 m long. However, all these studies are characterised
by a limited amount and types of monitoring data, that prevent a deep
understanding of the ongoing phenomena.
The present case study concerns the Val di Sambro (hereinafter, in-
dicated as VDS, in brief) twin tunnels, excavated in a geomorphological
unstable setting in the Northern Apennine (Emilia Romagna, Italy).
Indeed, over 70,000 landslides bodies have been identified in Emilia–
Romagna region (Bertolini, 2010), most of which are complex landslides,
⁎ Corresponding author.
structurally complex formations (Esu, 1977; D'Elia et al., 1998). Ac-
cording to the regional Inventory Map (Bertolini and Pellegrini, 2001)
one-third of them are active, while the remaining two-thirds are
dormant, but they can be reactivated by triggering factors, such as
snowmelt, earthquakes and intense and/or prolonged precipitations
(Carboni et al., 2001).
The rock-mass crossed by the VDS tunnels is a structurally complex
formation and is characterised by extremely slow landslide bodies
(sensu Cruden and Varnes, 1996) of various areal sizes but whose
depth was not well known before tunnelling. With the start of excava-
tion, an acceleration of the slope movements and light damage to
buildings in the neighbouring inhabited areas were observed in some
sections of the tunnel route, especially near the tunnel entrances. This
scenario, very complex and difficult to outline with simplified models,
required to set up a structured monitoring system in order to evaluate
the stability conditions of the area affected by the construction works.
After describing the main technical features of the tunnels, the geo-
logical setting and the geotechnical site conditions, the paper is basically
aimed at illustrating and commenting the most significant outcomes
of the monitoring system in the Southern zone of the tunnels, where
the major deformative phenomena occurred. In this study, surface
(topographic measurements), subsurface (inclinometers, piezometers)
and underground (convergence measurements) conventional moni-
toring data, together with space-borne Synthetic Aperture Radar inter-
ferometry (InSAR) data, were interpreted to analyse the magnitude

http://dx.doi.org/10.1016/j.enggeo.2015.07.001
0013-7952/© 2015 Elsevier B.V. All rights reserved.
72 A. Bandini et al. / Engineering Geology 196 (2015) 71–87

Fig. 1. Location of the “Valico Variant” project and of the VDS tunnels (modified from Autostrade per l'Italia, 2006). On the right, a picture of the Southern entrances of the two tunnels.

and the rate variation of the slope movements before, during and after
tunnelling.
The combined use of different monitoring techniques and the large
amount of acquired data allowed assessing the main features of the
landslide movements in undisturbed conditions and their relation
with the tunnelling activities. In particular, thanks to the ESA ERS ar-
chives of images gathering data since the early Nineties, satellite data
were very useful to estimate the surface ground movements well before
tunnelling, integrating the information provided by the traditional
monitoring techniques, here implemented only after the start of tunnel
excavation.
In recent years, thanks to the development of Multi Temporal Inter-
ferometry methods, as the Persistent Scatterers Interferometry (PSI)
(Ferretti et al., 2001), the InSAR has increasingly shown its usefulness

Fig. 2. Map of the area under investigation in proximity of the Southern entrances of the VDS tunnels. The location of the vertical and horizontal boreholes and the sections B–B and N–N,
represented in the Figs. 12 and 13, are also shown.
 A. Bandini et al. / Engineering Geology 196 (2015) 71–87
Fig. 3. Advancement versus time of the two tunnel faces and corresponding values of the GSI estimated at the tunnel faces according to Marinos and Hoek (2000).
in long-term monitoring of extremely slow landslides (Bovenga et al.,
2006; Cascini et al., 2009; Righini et al., 2012; Del Ventisette et al.,
2014; Wasowski and Bovenga, 2014), especially when combined with
in situ surveying, ground observations, topographic and geotechnical
measurements (Colesanti et al., 2003; Hilley et al., 2004; Colesanti and
Wasowski, 2006; Calò et al., 2012; Tofani et al., 2013).
2. Tunnel description and site conditions
The VDS project is part of a huge program of improvements of the
Italian A1 Milan–Naples highway in the mountainous region between
Bologna and Florence (Fig. 1). This program involves a segment of the
motorway approximately 60 km long. It consists partly in widening
the current carriageway by adding a third lane with emergency hard
shoulders in each traffic direction (“Third lane” project) and partly
in straightening the highway axis by constructing new stretches (“The
Valico Variant” project). A total length of about 29 km of the “Valico
Variant” are in underground and the VDS tunnel is the second longest
one. As the other underground sectors of the “Valico Variant”, it consists
of twin tunnels (Fig. 2), identified, in the present paper, as “North
tunnel” (for the traffic towards Bologna) and “South tunnel” (for the
traffic towards Florence).
Each tunnel, whose excavation was completed in early November
2014, is 3.8 km long and has a section of approximately 180 m2 with a
maximum span of 16 m. It will include two traffic lanes, 3.75 m wide,
and a security lane with the same width. The distance between the
tunnel axes varies from 30 m to 35 m.
The two tunnels were excavated into a hillside extending from an
elevation of almost 780 m a.s.l. (watershed) to approximately 340 m
a.s.l. (Setta torrent valley). The overburden above the tunnels ranges
Fig. 4. Images of core boxes of: a) sandstone rock material and b) pelitic material.
73
from a few metres at the entrances until a maximum value of 125 m
(North tunnel) and 116 m (South tunnel). They were driven with
the full-face method using traditional techniques and with fibreglass
dowel reinforcement of the face. The primary lining was made of steel
ribs with a shotcrete layer, with different characteristics according to
the local geomechanical conditions. The permanent concrete lining, re-
inforced if necessary and completed with the casting of a tunnel invert,
followed at a distance between one and eight tunnel diametres or im-
mediately behind the face where it was required to promptly reduce
the recognised deformation phenomena of the tunnel section.
In the southern zone of the tunnels (Fig. 2), examined in the present
paper, the excavation started in February 2010 in the North tunnel and
four months later in the South one (Fig. 3). The advance rate was always
less than 1.5 m/d with lower values (about 0.6–0.7 m/d) in the entrance
area where some instability phenomena occurred and particularly low
values of the Geological Strength Index (GSI) (Hoek, 1994; Hoek and
Brown, 1998) were estimated. The excavation proceeded also very
slowly at a distance from the entrance ranging from 100 to 400 m
(Fig. 3) for the occurrence of marked landslide movements, which de-
termined damage to the final lining. In addition, tunnelling stopped
for a short period during August 2011 and for a long period between
January and May 2012 (Fig. 3), in order to understand if the observed
acceleration of the landslides movements, described in the paper,
was induced by tunnelling and to identify the excavation parameters
affecting the phenomenon.
Most of the tunnel route crosses the Monghidoro unit (hereinafter
indicated as MOH), a structurally complex formation from a geotechni-
cal point of view (Esu, 1977; D'Elia et al., 1998). Indeed, it is a heteroge-
neous rock-mass (Carboni et al., 2001; Borgatti et al., 2006) with
variable percentages of sandstone (Fig. 4a) and silty clay (Fig. 4b).
74 A. Bandini et al. / Engineering Geology 196 (2015) 71–87

Fig. 5. Image and pictorial sketch of the tunnel face characterised by mixed soil conditions.

According to the prevailing component (engineering rock or soil), are-
naceous turbidites with different thickness up to isolated sandstone
blocks embedded into a silty-clay matrix (Fig. 5) may be encountered
(Picarelli et al., 2010). Consequently, the MOH can be locally assimilated
also to bimrocks according to the Medley definition (1994). The sand-
stone is characterised by a medium to fine grain size distribution, at
times coarse or micro-conglomeratic. In the rock-mass volumes where
the highly fractured sandstones prevail, the joints, especially in the
coarse fraction, are generally weakly cemented. Large volumes of
the MOH are brecciated and tectonised, giving a chaotic structure to
the rock-mass, as clearly visible on the tunnel face during tunnelling
(Fig. 5).
The hillslope crossed by the twin tunnels is affected by complex
landslide bodies of various areal sizes, which were prone to instabil-
ities already in the past. For instance, when the A1 highway was
built in the Sixties, a stretch of the motorway next to the VDS tunnel
route was abandoned and moved into underground because of land-
slide movements. Afterwards, some years before starting the excava-
tion, a geostructural study (Crosta et al., 2003) was performed to
identify and to delimit the boundaries of the landslide bodies in the
zone between La Quercia and Badia Nuova (Fig. 1). The study was per-
formed on a wide area, long approximately 17 km and with a mean
width of 4 km around the Setta torrent valley, through photogramme-
try followed by an in situ survey. The resulting inventory map highlighted
a situation of widespread instability in the investigated area, characterised
by diffuse landslide bodies of great areal extension within which small-
er instabilities are often recognisable, both on the left and right bank of
the Setta torrent. Near the Southern entrances of the tunnels, Crosta
et al. (2003) identified complex landslides composed by large main
bodies, classified as relict, with secondary dormant slides inside
(Fig. 6), without however providing any information about their exten-
sion at depth.
Fig. 7 represents the landslide bodies recognised in the area, soon
after the start of the tunnel excavation, by Picarelli and Viggiani
(2012) through in situ observations of terracing and trenches.
The geotechnical characterisation of the MOH and of the more su-
perficial landslide bodies was extremely difficult, due to their chaotic
nature. Table 1 summarises the values of the mechanical parameters
used in the design (Autostrade per l'Italia, 2007). The strength values
of the landslide materials were obtained after a back-analysis of the

Fig. 6. Extract of the landslide inventory map near the VDS tunnel entrance (modified from Crosta et al., 2003). The sections B–B and N–N, represented in the Figs. 12 and 13, are also drawn.
 A. Bandini et al. / Engineering Geology 196 (2015) 71–87
Fig. 7. View of the slope in which the tunnel is excavated (modified from Picarelli and Viggiani, 2012). The perimeters of the main landslide bodies identified by geomorphologic studies are
also shown.
landslide movements near the tunnel entrances, where an increase in
the displacement rate was noted after the tunnelling start. The MOH
was conversely characterised using the approach proposed by Marinos
and Hoek (2000) for flyshoid formations.
The hydraulic conditions in the rock-mass are significantly affected
by the heterogeneity of the MOH and the permeability is mainly related
to the fracturing state of the sandstone layers, due to the low intrinsic
permeability of the sandstone and pelitic components.
A synthesis of the GSI values estimated at the tunnel faces during the
construction is provided in Fig. 3. Very low values, ranging between 25
and 30, were identified near the entrance, while larger values, even
higher than 45, were observed gradually moving away towards North
from a chainage of 1000 m, due to the increment in the arenaceous per-
centage of the MOH formation.
3. Slope monitoring
The stability of the area affected by the tunnel excavation was ex-
tensively studied with surface and subsurface investigations through
an integrated monitoring system. The subsurface monitoring was
carried out with conventional geotechnical methods (inclinometers,
Casagrande piezometers), while the surface movements were con-
trolled with a robotic topographic network and with SAR interferomet-
ric techniques.
The intense vegetation cover, together with the limited number of
buildings and infrastructures, and the rate of the slope movements in
the investigated area greatly affected the ground investigations and
the subsurface measurements. Consequently, the monitoring system ar-
chitecture was modified several times during the tunnel construction
on the basis of the measured data, as described below.
Table 1
Main mechanical properties of the different geomechanical units (E: Young's modulus;
c': effective cohesion; ϕ': effective friction angle) (Autostrade per l'Italia, 2007).
E [MPa] c' [kPa]
Landslide deposits 20 + 2.5 z(⁎) 10
Fractured MOH 500 75
MOH 1200–3450 100–375
z*: depth below the ground level.
75
3.1. Subsurface monitoring
The position of the inclinometers and Casagrande piezometers is
shown in Fig. 8. It can be noted that they are grouped in two different
regions: most of them are located near the tunnel entrances, inside
the landslide body SB2 of Fig. 7, while a smaller group is placed in the
area undercrossed by the tunnels at a distance from the entrance
between 400 and 700 m, inside the landslide body SD of Fig. 7.
The Casagrande cells were installed in boreholes, typically drilled
close to the inclinometers, up to a depth of 51 m b.g.l. The evolution in
time of the water depth measured by the piezometers is reported in
Fig. 9. As it can be observed, no measurements were performed before
the start of the tunnel construction, as they began in July 2010. After
an initial transient, they fluctuate seasonally around a mean value
over the observation period. The maximum values are recorded, every
year, at the beginning of the spring, while a decrease is observed after
April in most of the instruments. The oscillations in the water level are
more significant in those piezometers indicating a deeper water table,
suggesting the presence at depth of high permeability zones, probably
associated to the more intense fracturing of the MOH.
As concerns the inclinometric monitoring, 14 inclinometric tubes
were initially installed in 2010, near the tunnel entrances in the land-
slide body SB2, followed by other 5 inclinometers in the landslide body
SD in the next years. As it can be noticed in Fig. 8, they are mainly
located near the two tunnels, especially the ones at the entrance,
with a planimetric distance to the tunnel axes ranging from a few
metres to 280 m (inclinometer 14) and a depth varying between 34 m
(inclinometer 6) to 97 m b.g.l. (inclinometer 16). The measurements
were initiated before the start of tunnelling only in 5% of the inclinom-
eters, making possible to follow the evolution of the movements also
prior and during the tunnel face passage. Indeed, due to the delay in
the installation, most instruments allowed only to study the magnitude
and the rate of the displacements while the tunnel faces gradually
moved away. In addition, the monitoring period was very short in
some boreholes, as the inclinometric tubes broke soon after their instal-
lation for the relevant recorded displacements. Near the dismissed incli-
ϕ' [°] nometers, new boreholes (indicated with code “bis” in Fig. 8), deeper
30 than the previous ones, were subsequently drilled and equipped with
30 inclinometer tubes to substitute the ones no longer usable. Consequent-
28–32 ly, the inclinometer number has been increased over time by 50% as
compared to the initial monitoring plan.
76 A. Bandini et al. / Engineering Geology 196 (2015) 71–87
Fig. 8. Location of the piezometric and inclinometric boreholes. The vectors represent the cumulated inclinometric displacements at the ground surface associated to the advancement of
the two tunnel faces over periods of 6 months.
The inclinometric measurements were performed almost weekly
with readings at intervals of 0.6 m along the length of the guide tubing.
Fig. 10 shows an example of typical cumulative displacements versus
depth profiles relative to the inclinometric tube 11, long 59 m. As it
can be seen in Fig. 10, the movements predominantly occur along
multiple shear surfaces that slide on each other along the direction
of the local maximum slope gradient (Fig. 11). At the various sliding
surfaces, the displacement vector typically undergoes a slight rotation,
represented by the change in the azimuth angle in Fig. 11. The cumula-
tive displacement profiles of Fig. 10 indicate two main discontinuities at
the depths of 33 m and about 57 m b.g.l.. The magnification box in
Fig. 10 shows significant displacements near the borehole bottom,
which reached 65 mm 7 months after the tube installation at the
depth of 56 m. The most relevant displacements are observed in corre-
spondence of the sliding surface at 33 m b.g.l., where the tube rupture
occurred 10 months after its installation.
A comprehensive analysis of the integral inclinometric move-
ments highlights that the depth of the shallowest sliding surfaces
ranges from 10 to 30 m b.g.l., while the deepest ones are found
also over 80 m b.g.l., as suggested by the displacements measured
at the bottom of the tubes by the majority of the instruments, in-
cluded the “bis” ones. Unfortunately, it was not possible to establish
the exact depth of the deepest sliding surfaces because of the limited
length of the inclinometric boreholes. Since the bottom of the incli-
nometer tubes was often not fixed, the inclinometric measurements
did not permit to determine the real displacement magnitude at the
sliding surfaces and at the ground surface, which is consequently
underestimated.
The investigated stretch of the two tunnels is generally included in
the sliding masses. Fig. 12 shows that along the section B–B inside the
landslide body SB2 near the tunnel entrances (Fig. 2) the deepest sliding
surface is identified at a maximum depth of about 65 m b.g.l. and it under-
crosses the two tunnels that at this chainage were excavated between 40
and 50 m b.g.l. Furthermore, the downhole displacements (zoom boxes
of the inclinometric profiles in Fig. 12) possibly indicate creep phenomena
at depth. Within the main deep landslide body, secondary shallow slides at
maximum depths of 30–35 m b.g.l. can be recognised. This reconstruction
confirms the model of complex landslides with compound slide move-
ments proposed by Crosta et al. (2003).
Along the section N–N drawn inside the landslide body SD (Fig. 13),
movements concentrate along a sliding surface having a maximum
depth of approximately 20 m b.g.l., well above the tunnel position. How-
ever, also in this case, the displacements measured at the tube bottom
even in the longest inclinometric borehole (inclinometer 16, long
97 m) support the presence of deeper movements affecting the portion
of the boreholes below the sliding surface.
In order to study the existence of a relationship between the excava-
tion advance and the slope movements, the influence of the inclinome-
ter distance to the tunnel face was examined. In particular, for each
inclinometer, a reference chainage was determined projecting the in-
strument position on the North tunnel axis. The horizontal axis in
Fig. 14 represents the distance between this reference chainage and
the position of the tunnel face. The zero abscissa thus corresponds
to the face passage at the reference section and, in other words, to
the minimum distance between the inclinometer and the North tun-
nel face. The negative values correspond to positions of the face
 A. Bandini et al. / Engineering Geology 196 (2015) 71–87
Fig. 9. Measurements of the piezometric head by Casagrande piezometers (a) and location of the instruments in the investigated area (b) (the depth of the Casagrande cells is indicated
into parenthesis in the legend).
approaching the inclinometer, while the positive ones indicate that
the face is moving away.
The example of Fig. 14, referring to the inclinometer 11, shows that
at a depth of 33 m, the displacement rate increases when the tunnel
face is approaching the inclinometer and decreases when it is moving
away, with a roughly symmetrical profile with respect to the “zero
distance”. Conversely, the displacement rate on the deeper sliding sur-
face, at 57 m b.g.l., increases when the tunnel face is approaching,
reaches the maximum value at the minimum distance and remains con-
stant in the observation period, when the tunnel face is moving away.
It should be noted that the number of inclinometers that allow this
analysis is limited, because the maximum distance between each incli-
nometer and the tunnel face in the approaching and moving away
phases is fairly reduced. Therefore, on the basis of these considerations,
the analysis cannot be completely generalised.
The recording period of inclinometer 4, very close to the North
tunnel, whose monitoring started two months before the advancing
face passage, is characterised by three temporary excavation stops
(Fig. 15). The differential displacement rate along the shallow sliding
surface (i.e. 20 m b.g.l.) decreases almost imperceptibly during the
first stop occurred when the tunnel face was approaching the moni-
tored point, while it reduces of one order of magnitude during the
77
second stop occurred when the face was moving away (Fig. 15). The
abscissa scale in the plot of Fig. 15 does not allow highlighting the
bell-shape trend of Fig. 14a, because soon after the face passage
the inclinometric tube broke in correspondence of the sliding surface.
It can be noted that, in this case, as for the other inclinometers very
close to the tunnels, the effects of the subsidence induced by the excava-
tion on the displacements may be not negligible, but in the present
paper they were not analysed.
When the tunnel face reaches large distances from the monitored
area, the slope movements seem no more influenced by the excavation
works. As an example, in Fig. 16, relative to two inclinometers located in
the SB2 (inclinometer 12) and SD (inclinometer 19) landslide bodies, the
differential displacement rate on the shallow sliding surface (located in
both cases at about 19 m b.g.l.) significantly drops with the progressive
tunnel face advancement. The two trends are completely similar, but,
because of a different minimum distance of the two inclinometers to
the tunnel (130 m for the former and 60 m for the latter), the maximum
rate values observed with the inclinometer 19 are three times higher
than those recorded by the other one.
The differential displacement rate measured in the borehole 12 indi-
cates an influence of the piezometric conditions on the movements. In
fact, the values show that the displacements accelerate during the
78 A. Bandini et al. / Engineering Geology 196 (2015) 71–87

Fig. 10. Example of inclinometer measurements (inclinometer 11). The cumulative displacement versus depth profiles show the presence of movements also at the borehole bottom.

spring of the years 2013 and 2014, in correspondence of the increase in
the groundwater level recorded by the close piezometer (Fig. 9). Fig. 17,
referring to the borehole 8 near the Southern tunnel entrances (Pz_08 in
Fig. 9), further highlights the influence of the hydraulic conditions on
the slide movements. The local increments in the movement rate on
the sliding surface at a depth of about 19 m b.g.l. can be clearly associat-
ed to the raise in the water level. As the tunnel face moves away, the
31.2 m deep discontinuity was no longer distinguishable.
The whole representation of the inclinometric measurements
over time is reported in Fig. 18 in terms of differential displacements
along the different sliding surfaces versus time. The inclinometric mea-
surements inside the landslide body SB2 and SD are depicted with differ-
ent colours taking into account the depth of the sliding surfaces. The
position of the two tunnel faces in time is also indicated for sake of com-
parison. The considerations already discussed for single inclinometers
are clearly valid for the total set of instruments. In addition, Fig. 18
shows the effect of the long-term excavation stop of Fig. 3, which deter-
mined a temporary deceleration of the movements.
The corresponding rate of the differential displacements displayed
in Fig. 18 are presented in Fig. 19 adopting the same colour legend.
The values are provided in terms of distance between the inclinometers
and the position of the North tunnel face, already clarified. The maxi-
mum rates were recorded for inclinometer — face distances ranging
from −50 m to 100 m, with the highest values recorded at the shallow
sliding surfaces, especially in the landslide body SB2. For distances larger
than 300 m the movement rates tend to concentrate below a threshold

Fig. 11. Variation with depth in the azimuth angles of the inclinometric displacements: inclinometer 7(a), downstream of the tunnels, and 11bis (b), in the upstream side.
 A. Bandini et al. / Engineering Geology 196 (2015) 71–87 79

Fig. 12. Interpretation of the inclinometric measurements along the Section B–B of Fig. 2, identifying sliding surfaces at different depths.

value (approximately 5 mm/month) in both the landslide bodies SB2
and SD, with some exceptions related to the seasonal increments due
to water table oscillations. These values can be considered as represen-
tative of the landslide natural activity.
3.2. Ground monitoring through conventional and SqueeSAR techniques
In Fig. 8 the inclinometric cumulative displacements at the ground
surface are represented by segments of length proportional to the
movements recorded during successive periods of 6 months. They are
correlated with the tunnel face advances in the same periods. It is evi-
dent that the tunnel passage induces large movements at the ground
surface, that reach, in some cases, a few tens of centimetres on the
whole, such as for the inclinometers 10 and 11, characterised by maxi-
mum displacement rate of about 40 mm/month. These movements con-
tinue also when the faces are distant, even if they are less marked, and
follow a downward trend approximately in the direction of the maxi-
mum slope gradient.
Topographic measurements were carried out with automatic total
stations (accuracy of 1 mm in distance and 1 arcsec in angle) on the in-
clinometer heads near the tunnel entrances. Fig. 20 illustrates the tem-
poral evolution of the horizontal displacements of some representative
targets. The measurements started when the tunnel faces were already
distant but, however, they indicate some correlations with the excava-
tion activities. In particular, they show a marked tendency to movement
stabilisation in time. The comparison between the topographic and
the corresponding inclinometric measurements was possible only for
one instrument and it highlights an agreement in terms of recorded hor-
izontal displacements at the ground surface. Also in this case, the hy-
draulic conditions are found to affect the displacement rate especially
in the last two years of monitoring, when the influence of the excavation
works can be considered negligible.

Fig. 13. Interpretation of the inclinometric measurements along the Section N–N of Fig. 2. Some integral displacements versus depth profiles highlight movements also at the inclinometric
borehole bottom.
80 A. Bandini et al. / Engineering Geology 196 (2015) 71–87

Fig. 14. Inclinometer 11: differential displacements and related rates on the two sliding surfaces at 33 m and 57 m b.g.l. as a function of the distance between the inclinometer reference
chainage and the face of the North tunnel.

Basically, the conventional monitoring described up to now mainly
began after the start of tunnel excavation. In order to investigate the
slope activity even before tunnelling, SAR interferometry analyses per-
formed in the area through the SqueeSAR technique (Ferretti et al.,
2011) were also considered. The SqueeSAR analysis undertaken by
Tele-Rilevamento Europa (T.R.E.) consisted of processing SAR images
acquired with a descending orbital geometry by the two satellites ERS
1–2 and RADARSAT-1 over an overall period of twenty years. The two
image datasets cover different time periods: the ERS data (80 images)
range from April 1992 to January 2001, thus providing information
about the slope movements before tunnelling, while the RADAR-SAT
data (106 images), going from March 2003 to April 2012, refer to obser-
vational windows prior and during the excavation. The descending ge-
ometry was chosen taking into account the orientation of the hillside
interested by tunnelling and of the maximum slope line, assumed as
the preferential direction for sliding phenomena.
The two image datasets were processed to compute the movement
magnitude and rate of natural targets available on the ground surface.
The algorithm provides the component of the displacement vector
along the line of sight (LOS), i.e. the line joining the radar and the target.

Fig. 16. Decrease in the inclinometer differential displacements on the sliding surface
Fig. 15. Inclinometer 4: effect of the temporary excavation stops on the differential dis- at approximately 19 m b.g.l.: comparison between two inclinometers located in the SB2
placement rate on the sliding surface at about 20 m b.g.l. (inclinometer 12) and SD (inclinometer 19) landslide bodies.
 A. Bandini et al. / Engineering Geology 196 (2015) 71–87 81

Fig. 17. Effect of the piezometric level on the differential displacements along the two sliding surfaces identified by inclinometer 8.

Fig. 18. Differential displacements on the shallow and deep sliding surfaces recorded by the inclinometers in the landslide bodies SB2 and SD.

Positive displacement values indicate movements towards the satellite
(shortening along the LOS), while negative displacements denote
movements away from the sensor (lengthening along the LOS).
In the area investigated through interferometry, that extends for
about 1.0 km from the Southern tunnel entrance to the North, the
natural targets were limited in number, because of the high vegetation
cover and the lack of buildings. In particular, approximately 60 and 35
stable points were identified in the ERS and R-SAT images, characterised
by coherence mean values of 0.72 and 0.56, respectively. They are con-
centrated mainly on the sides of the old A1 highway and on its

Fig. 19. Differential displacement rates on the shallow and deep sliding surfaces recorded by the inclinometers in the landslide bodies SB2 and SD.
82 A. Bandini et al. / Engineering Geology 196 (2015) 71–87

Fig. 20. Horizontal displacements of the topographic targets placed at the heads of the inclinometer boreholes.

carriageways, with an inhomogeneous distribution (Fig. 21). As it can be
seen in Fig. 21, where the permanent scatterers (PS) are coloured differ-
ently according to the mean displacement velocity along the LOS, signif-
icant movements were recorded also before tunnel excavation. Indeed,
in the monitoring period 1992–2001, the ground displacement rate
along the LOS reaches a maximum value of approximately 10 mm/
year (targets A015P and A015O on the old highway carriageways, left
picture in Fig. 21). The displacement rate, in most cases, is around 5–6
mm/year except for the targets in the upper part of the left picture of
Fig. 21, that practically do not move.
In order to compare the ground movement rate of the ERS and R-SAT
targets, the displacements along the LOS were projected on the horizon-
tal plane, because the images were acquired by the two satellites with
different off-nadir angles (24.1° for ERS and 34.3° for R-SAT). Only in
a few cases, there is a correspondence between the positions of the co-
herent points obtained from the two image datasets. As an example, the
displacements on the horizontal plane versus time of three representa-
tive ERS permanent scatterers are compared with the trends of the
neighbouring R-SAT points in Fig. 22. In all cases, the movements do
not increase regularly over time and they show continuous changes in
the displacement rate with a strong data scatter, which hides the local
rate variations. The movements of the ERS points exhibit a roughly cur-
vilinear trend with a progressive velocity reduction especially from
1999. Generally, before the end of November 2010 the displacement
rate of the targets is comparable, irrespective of the dataset. After the
North tunnel excavation start, the displacements of two (000JI and
000JJ) of the three R-SAT points in Fig. 22 become more pronounced
with a movement acceleration in the last years. The mean displacement
rate significantly increases from 18 mm/year (000JI) and 13 mm/year
(000JJ) in the first period to 80 mm/year for both the targets in the latter
period starting from the end of November 2010. On that date, the North
tunnel face, the closer to the coherent points of the interferometry anal-
ysis, was about 400 m distant from the two permanent scatterers. In
April 2012 (end of the R-SAT monitoring period), the two targets
moved totally 260–300 mm (Fig. 22). Only few targets show a similar
piecewise linear trend characterised by a higher displacement rate

Fig. 21. Spatial distribution and mean LOS displacement rate of the coherent targets (ERS, on the left, and R-SAT, on the right) in the area surrounding the Southern entrance of the VDS
tunnels.
 A. Bandini et al. / Engineering Geology 196 (2015) 71–87 83

Fig. 22. Evolution in time of the horizontal displacements of six targets: comparison between ERS (on the left) and R-SAT (on the right) data.

during the tunnelling operations and they are all located in the hillside increasing in time, as the target 000JL in Fig. 22. In any case, any velocity
next to the old A1 highway. Most of the R-SAT points next to the tunnel trends differing by the linear increasing and the piecewise linear have
entrance in the period 2003–2012 moved with a rate regularly not been observed (Fig. 22).
84 A. Bandini et al. / Engineering Geology 196 (2015) 71–87
Fig. 23. North tunnel: time elapsed between the initial and final measurements (a) and distance between the face and the monitoring section at the time of the first and final readings (b).
4. Tunnel monitoring
The tunnel wall displacements were monitored with geodetic mea-
surements carried out on the primary and permanent linings. Measures
were performed with a total station (accuracy of 1 mm in distance and 1
arcsec in angle), which recorded, at different tunnel positions, the 3-D
(transversal, longitudinal and vertical) coordinates of the targets placed
along the tunnel profile with respect to a zero reading in a local refer-
ence system.
Almost 300 sections on the temporary lining of each tunnel were
monitored, at a mean spacing of 5–7 m. Each monitoring section in-
cluded five optical targets installed on the crown and on the sides of
the tunnel lining, according to the scheme of Fig. 23. The reflectors
were welded to bolts, anchored to the surrounding rock-mass, during
the installation of the steel ribs. Measurements were taken periodically,
usually once per day, as soon as the shotcrete spraying on the steel rib
and the positioning of the optical targets were completed, at a variable
distance from the face (Fig. 23), always larger than 1 m. Consequently,
the measured convergence values in the monitoring sections cannot
be directly compared one to the other, but can provide indications on
movements in progress from a qualitative point of view.
Generally, the measurements on the temporary lining covered little
periods of time (on average, 20 days, Fig. 23), because the permanent
lining followed at a short distance from the face. In each tunnel the
monitoring was extended to about 30 sections on the final lining in
order to check the long-term behaviour of the construction.
Fig. 24 shows the final average transversal displacements of two
diametrically opposite targets (1 and 5) on the sides of the primary
lining. These values present a high scatter, generally fluctuating around
zero. However, along a tunnel stretch at a distance of 100 m from the
entrances and long about 230 m (marked by the shaded areas in
Fig. 24), the values are always negative, thus indicating tunnel move-
ments towards the Setta river valley. In some sections of this moving
stretch of the two tunnels, the displacement vectors of the five targets
are all oriented downstream, as indicated in the example of Fig. 25(a)
with almost negligible movement components parallel to the tunnel
axes. Most sections show asymmetry in the deformation and appear
distorted (Fig. 25(b)). These observations are congruent with the iden-
tification in this area of the landslide body SB2 and with the indications
of the slope monitoring. It should be noted that, conversely, other sec-
tions of the tunnels are characterised by positive values of the average
transversal displacement (Fig. 24). This finding could be explained by
considering the possible non-fixed position of the reference points in
the topographic measurements. The larger positive values were gener-
ally recorded after significant stops in the excavation works. During
these periods, the higher transversal displacement values recorded by
the targets on the tunnel valley side compared to the month one can
be due to the translation of the tunnel axis (and, as a consequence,
of the measure reference points) towards valley in consequence of
the slope movements. If the variation of the tunnel axis were
accounted for, the average values of the transversal displacements
would become zero or assume also negative values, compatibly with
the landslide kinematics.
The movements tend to run out over time. The vertical (a) and trans-
versal (b) displacements of the two targets 1 and 2, placed on the left
side of the permanent lining in a section of the tunnel stretch moving to-
wards valley, are represented in Fig. 26. The displacement amount and
rate of the two targets vary with the distance between the tunnel face
and the monitoring section (Fig. 26(c)), but the movements tend to
reduce as the face moves away with negligible velocities at distances
larger than 400 m.
5. Discussion and final remarks
The paper presented the case study of the Val di Sambro large
section twin tunnels excavated in a structurally complex formation af-
fected by landslides of various areal sizes. For the large amount of mon-
itoring data collected during the excavation, it represents a unique case
history in the study of the interaction between tunnelling and land-
slides. In fact, the installed integrated monitoring plan, based on
 A. Bandini et al. / Engineering Geology 196 (2015) 71–87 85

Fig. 24. Final average transversal displacements of the two targets 1 and 5 at the sidewalls of the primary lining along the two tunnel routes (the negative sign indicates downstream
displacements towards the Setta river valley).

Fig. 25. Displacement vectors of the targets on the primary lining: comparison between two monitoring sections of the North (a) and South (b) tunnels in the stretch characterised by
movements towards the Setta river valley (section chainage: 268 m (a) and 173 m (b)).
86 A. Bandini et al. / Engineering Geology 196 (2015) 71–87

Fig. 26. Example of temporal evolution of the vertical (a) and transversal (b) displacements of the targets 1 and 2 (left sides) on the North tunnel final lining (section at chainage 118 m).
The distance of the considered monitoring section from the N tunnel face over time is also shown (c).

coupling traditional monitoring methods (inclinometers, piezometers,
underground and surface topographic measurements) with SAR inter-
ferometric technique, allowed to investigate the behaviour of the
slope in relation to the tunnel excavation and in the long term.
In particular, referring to the examined case study, the monitoring
provided interesting results which lead to the following considerations:
• the information deriving from inclinometers could help in defining
the location of the sliding surfaces and the in-depth development of
the sliding masses. Complex landslide bodies characterised by multi-
ple shear surfaces that slide on each other along the direction of the
locally maximum slope gradient were identified. The shallow sliding
surfaces are approximately located at a depth ranging from 10 to`
30 m b.g.l., while the deepest ones are found also over 80 m b.g.l. Un-
fortunately, the limited length of the inclinometric boreholes, mostly
lower than 60 m, did not allow investigating over such a depth;
• the two tunnels are generally included in the moving mass and inter-
sect the sliding surfaces of the landslide bodies SB2 and SD;
• the results of SqueeSAR analysis on ERS data indicate that extremely
slow landslide movements (Cruden and Varnes, 1996), with a mean
displacement rate of approximately 10 mm/y, were recorded even
before the tunnel excavation. A careful analysis shows that the move-
ments were characterised by continuous changes in the displacement
rate;
• the tunnelling caused a displacement acceleration along the sliding
surfaces at the various depths, reaching a maximum movement rate
of about 40 mm/month at the ground surface with the face passage.
The movements were particularly significant in a stretch of both the
tunnels, long approximately 230 m, extending from a distance to the
entrances of 100 m, as highlighted by the topographic measurements
on the temporary lining;
• the sliding surfaces, shallow and deep, display a different kinematic
behaviour, when the tunnel face is moving away, but this analysis
could not be generalised as, unfortunately, all measurements started
to be collected after the beginning of excavation works, thus not
allowing to have a long observation period of the slope behaviour in
undisturbed conditions;
• the slope movements continued in time even when the tunnel faces
were at a very large distance, but with much lower rates at distances
from the tunnel face larger than 200 m. The landslide movement
 A. Bandini et al. / Engineering Geology 196 (2015) 71–87
stabilisation over time is also confirmed by the topographic measure-
ments on the ground surface and on the tunnel final lining of both the
tunnels;
• for those inclinometers already very far from the tunnel faces, the
displacements, as expected, resulted to be well correlated to the pie-
zometric level, showing an increment in the movement rate during
the spring seasons.
This case demonstrates that the study of the interaction between
tunnelling and landslide phenomena in structurally complex forma-
tions is challenging and requires a careful design of an effective moni-
toring system that should have the following features:
• it should consist of different techniques. In particular, as shown in this
paper, the combined use of traditional and interferometric monitoring
allows having a more complete view of the long-term stability of the
area interested by tunnelling;
• it should be set up and start to acquire before the tunnel construction,
and continuously measure during and after the tunnel construction.
The measurements with the different techniques should be per-
formed over the same periods to allow a comparison;
• it should be constantly updated with the acquired data. Indeed, the
data interpretation should be done simultaneously with the excava-
tion to have the possibility to analyse the influence of various factors
(excavation parameters, geomechanical local conditions) on the
movement magnitude and rate.
Acknowledgements
The Authors would like to thank Spea Autostrade per l'Italia and
TOTO for having provided the monitoring data. We are also grateful to
Eng. Sara Bertolin for the assistance with this work.
References
Autostrade per l'Italia, 2006. Variante di Valico. Newsletter 1, 1–4 (http://www.
autostrade.it/variante-di-valico/newsletter-variante-di-valico/index.html).
Autostrade per l'Italia, 2007. Autostrada (A1): Milano – Napoli. Adeguamento del tratto di
attraversamento appenninico tra Sasso Marconi e Barberino di Mugello. Tratto: La
Quercia – Badia Nuova. Sub-tratta: Lagaro – Badia Nuova. Lotto 6–7. Progetto Esecutivo.
Galleria Val di Sambro. Relazione Geomeccanica.
Barla, G., Chiappone, A., Vai, L., 2006. Slope monitoring system. In: Barla, G., Barla, M.
(Eds.), Instabilità di versante. Interazioni con le infrastrutture, i centri abitati e
l'ambiente. XI Ciclo di Conferenze di Meccanica e Ingegneria delle Rocce (MIR
2006). Pàtron Editore, Bologna, pp. 177–202 (Cap.8).
Bertolini, G., 2010. Large earth flows in Emilia Romagna (Northern Apennines, Italy): or-
igin, reactivation and possible hazard assessment strategies. Z. Dtsch. Ges. Geowiss.
161 (2), 139–162. http://dx.doi.org/10.1127/1860-1804/2010/0161-0139.
Bertolini, G., Pellegrini, M., 2001. The landslides of the Emilia Apennines (northern Italy)
with reference to those which resumed activity in 1994–1999 period and required
Civil Protection interventions. Quad. Geol. Appl. 8, 27–74.
Borgatti, L., Corsini, A., Barbieri, M., Sartini, G., Truffelli, G., Caputo, G., Puglisi, C., 2006.
Large reactivated landslides in weak rock masses: a case study from the Northern Ap-
ennines (Italy). Landslides 3, 115–124. http://dx.doi.org/10.1007/s10346-005-0033-
9.
Bovenga, F., Nutricato, R., Refice, A., Wasowski, J., 2006. Application of multi-temporal dif-
ferential interferometry to slope instability detection in urban/peri-urban areas. Eng.
Geol. 88, 218–239. http://dx.doi.org/10.1016/j.enggeo.2006.09.015.
Bozzano, F., Cipriani, I., Mazzanti, P., Prestininzi, A., 2011. Displacement patterns of a land-
slide affected by human activities: insights from ground-based InSAR monitoring.
Nat. Hazards 59, 1377–1396. http://dx.doi.org/10.1007/s11069-011-9840-6.
Calò, F., Calcaterra, D., Iodice, A., Parise, M., Ramondini, M., 2012. Assessing the activity of a
large landslide in southern Italy by ground-monitoring and SAR interferometric tech-
niques. Int. J. Remote Sens. 33, 3512–3530. http://dx.doi.org/10.1080/01431161.
2011.630331.
Carboni, R., Casagli, N., Iotti, A., Monit, L., Tarchiani, U., 2001. La frana di Ca' di Sotto
(San Benedetto Val di Sambro, Bologna): caratteri geomorfologici, analisi geotecniche
ed interventi di mitigazione del rischio. Quad. Geol. Appl. 8, 93–106.
Cascini, L., Fornaro, G., Peduto, D., 2009. Analysis at medium scale of low-resolution
DInSAR data in slow-moving landslide-affected areas. ISPRS J. Photogramm. Remote
Sens. 64, 598–611. http://dx.doi.org/10.1016/j.isprsjprs.2009.05.003.
87
Colesanti, C., Wasowski, J., 2006. Investigating landslides with space-borne synthetic ap-
erture radar (SAR) interferometry. Eng. Geol. 88, 173–199. http://dx.doi.org/10.
1016/j.enggeo.2006.09.013.
Colesanti, C., Ferretti, A., Prati, C., Rocca, F., 2003. Monitoring landslides and tectonic mo-
tions with the Permanent Scatterers Technique. Eng. Geol. 68, 3–14. http://dx.doi.org/
10.1016/S0013-7952(02)00195-3.
Crosta, G.B., Cucchiaro, S., Siena, L., 2003. Autostrada (A1): Milano – Napoli Adeguamento
dell'attraversamento appenninico tra Sasso Marconi e Barberino del Mugello. Carta
Inventario dei fenomeni franosi nel tratto compreso tra La Quercia – Badia Nuova,
Lotti A – B. Dipartimento di Scienze Geologiche e Geotecnologie, Università degli
Studi di Milano Bicocca.
Cruden, D.M., Varnes, D.J., 1996. Landslides types and processes. In: Turner, A.K., Schuster,
R.L. (Eds.), Landslides: investigation and mitigationNational Research Council, Trans-
portation Research Board Special Report 247. National Academy Press, Washington
(DC), pp. 36–75.
D'Elia, B., Picarelli, L., Leroueil, S., Vaunat, J., 1998. Geotechnical characterisation of slope
movements in structurally complex clay soils and stiff jointed clays. Riv. Ital. Geotec.
XXXIII, 5–32.
Del Ventisette, C., Righini, G., Moretti, S., Casagli, N., 2014. Multitemporal landslides inven-
tory map updating using spaceborne SAR analysis. Int. J. Appl. Earth Obs. Geoinf. 30,
238–246. http://dx.doi.org/10.1016/j.jag.2014.02.008.
Esu, F., 1977. Behaviour of slopes in structurally complex formations. Int. Symp. The
Geotechnics of Structurally Complex Formations, Capri (Italy) vol. II, pp. 157–169.
Ferretti, A., Prati, C., Rocca, F., 2001. Permanent Scatterers in SAR Interferometry. IEEE
Trans. Geosci. Remote Sens. 39, 8–20. http://dx.doi.org/10.1109/36.898661.
Ferretti, A., Fumagalli, A., Novali, F., Prati, C., Rocca, F., Rucci, A., 2011. A new algorithm for
processing interferometric data-stacks: SqueeSAR. IEEE Trans. Geosci. Remote Sens.
49, 3460–3470. http://dx.doi.org/10.1109/TGRS.2011.2124465.
Hilley, G.E., Burgmann, R., Ferretti, A., Novali, F., Rocca, F., 2004. Dynamics of slow-moving
landslides from permanent scatterer analysis. Science 304, 1952–1955. http://dx.doi.
org/10.1126/science.1098821.
Hoek, E., 1994. Strength of rock and rock masses. ISRM News J. 2, 4–16.
Hoek, E., Brown, E.T., 1998. Practical estimates of rock mass strength. Int. J. Rock Mech.
Min. Sci. 34, 1165–1186. http://dx.doi.org/10.1016/S1365-1609(97)80069-X.
Jiao, Y.Y., Wang, Z.H., Wang, X.Z., Adoko, A.C., Yang, Z.X., 2013. Stability assessment of an
ancient landslide crossed by two coal mine tunnels. Eng. Geol. 159, 36–44. http://dx.
doi.org/10.1016/j.enggeo.2013.03.021.
Maiorano, R.M.S., Russo, G., Viggiani, C., 2014. A landslide in stiff, intact clay. Acta Geotech.
9, 817–829. http://dx.doi.org/10.1007/s11440-013-0249-0.
Marinos, P., Hoek, E., 2000. GSI: a geologically friendly tool for rock mass strength estima-
tion. Proc. GeoEng2000 Conference, Melbourne (Australia) vol. 1. Technomic Publish-
ing Co. Inc., Pennsylvania, pp. 1422–1440.
Medley, E.W., 1994. Engineering characterization of mélanges and similar block-in-matrix
rocks (Bimrocks) PhD dissertation Dept. Civil Engineering, Univ. of California at
Berkeley. UMI Diss. Svc., Ann Arbor, NI.
Noferini, L., Pieraccini, M., Mecatti, D., Macaluso, G., Atzeni, C., Mantovani, M., Marcato, G.,
Pasuto, A., Silvano, S., Tagliavini, F., 2007. Using GB-SAR technique to monitor slow
moving landslide. Eng. Geol. 95, 88–98. http://dx.doi.org/10.1016/j.enggeo.2007.09.
002.
Picarelli, L., Viggiani, C., 2012. Procedimento Penale n. 10488/2011 RG NR mod 44 “Frana
di Ripoli”. Valutazioni dei Consulenti di Parte della Società San Benedetto Val di
Sambro s.c.a.r.l. e del Raggruppamento Temporaneo di Imprese Vianini – Toto –
Profacta.
Picarelli, L., Petrazzuoli, S.M., Warren, C.D., 2002. Interazione tra gallerie e versanti. In:
Barla, G., Barla, M. (Eds.), Le opere in Sotterraneo ed il Rapporto con l'Ambiente IX°
Ciclo di Conferenze di Meccanica delle Rocce (MIR 2002). Patron Editore, Torino,
pp. 219–248.
Picarelli, L., Santo, A., Viggiani, C., 2010. Galleria Val di Sambro: Valutazione sulle
condizioni di stabilità del versante in corrispondenza dell'imbocco Sud.
Righini, G., Pancioli, V., Casagli, N., 2012. Updating landslide inventory maps using Persis-
tent Scatterer Interferometry (PSI). Int. J. Remote Sens. 33, 2068–2096. http://dx.doi.
org/10.1080/01431161.2011.605087.
Sciotti, A., Calabresi, G., 2000. Instability phenomena and remedial measures for a land-
slide in clay shales. Proc. 8th Int. Symp. On Landslides, Cardiff (UK) vol. 3,
pp. 1347–1352.
Sciotti, A., Pigorini, A., 2013. Interazione gallerie-versanti: “back-analysis” del dissesto
della galleria ferroviaria Scianina-Tracoccia. Gallerie e grandi opere in sotterraneo
108 pp. 49–52.
Tofani, V., Raspini, F., Catani, F., Casagli, N., 2013. Persistent scatterer interferometry (PSI)
technique for landslide characterization and monitoring. Remote Sens. 5, 1045–1065.
http://dx.doi.org/10.3390/rs5031045.
Urcioli, G., Picarelli, L., 2008. Interaction between landslides and man-made works. Proc.
10th Int Symp on Landslides, Xian vol. 2, pp. 1301–1307.
Wang, T.T., 2010. Characterizing crack patterns on tunnel linings associated with shear
deformation induced by instability of neighboring slopes. Eng. Geol. 115, 80–95.
http://dx.doi.org/10.1016/j.enggeo.2010.06.010.
Wasowski, J., Bovenga, F., 2014. Investigating landslides and unstable slopes with satellite
Multi Temporal Interferometry: Current issues and future perspectives. Eng. Geol.
174, 103–138. http://dx.doi.org/10.1016/j.enggeo.2014.03.003.