You are on page 1of 19

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/209803517

Measurement of landslide displacements using a wire extensometer

Article  in  Engineering Geology · February 2000


DOI: 10.1016/S0013-7952(99)00086-1

CITATIONS READS
64 1,252

7 authors, including:

Jordi Corominas José Moya


Universitat Politècnica de Catalunya Universitat Politècnica de Catalunya
148 PUBLICATIONS   4,514 CITATIONS    51 PUBLICATIONS   1,191 CITATIONS   

SEE PROFILE SEE PROFILE

Antonio Lloret Josep A. Gili


Universitat Politècnica de Catalunya Universitat Politècnica de Catalunya
115 PUBLICATIONS   2,516 CITATIONS    55 PUBLICATIONS   815 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Proyecto PROGRESS-CEE View project

Slope Mass-wasting under climate change. Physical mechanisms, predictive modelling and possible mitigation strategies View project

All content following this page was uploaded by José Moya on 18 January 2018.

The user has requested enhancement of the downloaded file.


Engineering Geology 55 (2000) 149–166
www.elsevier.nl/locate/enggeo

Measurement of landslide displacements using


a wire extensometer
J. Corominas a, *, J. Moya a, A. Lloret a, J.A. Gili a, M.G. Angeli b, A. Pasuto c,
S. Silvano c
a Dpto. Ingenierı́a del Terreno, E.T.S. Ingenieros de Caminos, Universitat Politècnica de Catalunya,
Jordi Girona 1–3. Edificio D-2, 08034 Barcelona, Spain
b IRPI-Consiglio Nazionale delle Ricerche, Via Madonna Alta 126, 06128 Perugia, Italy
c IRPI-Consiglio Nazionale delle Ricerche, Corso Stati Uniti 4. 35100 Padua, Italy

Received 21 April 1998; accepted for publication 15 April 1999

Abstract

A continuous recording of landslide displacements is often required in order to better understand the complex
relationship between the triggering factors and the dynamics of the movement. In this paper, we discuss the
performance of the borehole wire extensometer and the interpretation of its results. The analysis for the case of a
translational slide shows that the displacements measured with the wire extensometer are systematically smaller than
the movements observed at the ground surface. A relationship between the wire readings and the horizontal component
of the landslide movement has been established by means of three equations representing different stages of the wire
displacement within the borehole. The applicability of these equations and the interpretation of the wire extensometer
readings have been successfully checked at two landslide sites: Vallcebre in the eastern Pyrenees and Alverà in the
Dolomites. © 2000 Elsevier Science B.V. All rights reserved.

Keywords: Landslides; Monitoring; Pyrenees and Dolomites; Wire extensometer

1. Introduction and for understanding the dynamics of the process


of landsliding.
Reactivation or acceleration of landslides due The measurement of displacements of relatively
to rainfall events may occur in a very short period fast-moving landslides, often has many constraints.
of time. In these cases, observations made shortly Surveying of superficial ground movements using
after the event may miss significant information theodolites and Electronic Distance Measurement
relative to the triggering conditions of the move- ( EDM ) instruments requires a system of local
ments. Continuous recording of parameters such benchmarks that has to remain stable during the
as rainfall, pore water pressures or landslide dis- course of the investigation and to guarantee the
placements have become indispensable sources of visibility of the monuments. This surveying often
data for the analysis of the stability of the slope implies tedious and expensive campaigns of moni-
toring. Downhole monitoring devices like the
probe inclinometer do not allow continuous
* Corresponding author. Fax:+34-93-401-7251. recording of the displacements. Furthermore, due
E-mail address: corominas@etseccpb.upc.es (J. Corominas) to the failure of the borehole casing, probe incli-

0013-7952/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved.
PII: S0 0 1 3 -7 9 5 2 ( 9 9 ) 0 0 08 6 - 1
150 J. Corominas et al. / Engineering Geology 55 (2000) 149–166

nometers cannot work after displacements of only of a counterweight (Fig. 1). In order to protect
a few centimetres. However, the interpretation of the wire against corrosion, it is inserted into a
the intermittent measurements of the landslide plastic sheath and sealed with silicone. The move-
movement is rather difficult. For instance, the rate ment of the landslide displaces the wire either
of displacement between consecutive measures is inside or outside the borehole with the consequent
normally assumed as constant. This assumption, rotation of the pulley. The amount of wire dis-
however, may lead to a fairly inaccurate interpreta- placement is measured by means of a potentiome-
tion of the actual landslide behaviour because ter mounted on the axle of the pulley, which
many movements observe periods with a sudden provides an electrical signal proportional to the
acceleration, followed by lapses of quietness. angle of rotation. Readings of the potentiometer
Under such circumstances, the relations between may be carried out at defined time steps and the
displacements and ground water table fluctuations signal transmitted to a data logger. Long-life bat-
or rainfall records cannot be properly established. teries allow data storage for 2 or 3 months, and
To overcome these shortcomings, in the last few the device may last working for several metres of
years, new equipment allowing continuous meas- displacement, provided that enough wire has been
urements has been developed and improved. An left outside the borehole.
example of such equipment is the wire extensome-
ter that may be installed on either the landslide
surface (Bonnard and Steinmann, 1990) or into a
borehole (Angeli et al. 1988, 1989). The aim of 3. Interpretation of measurements
the paper is to present the performance of a wire
extensometer equipment placed within a borehole Wire extensometer readings cannot be properly
for monitoring landslide displacements. The char- interpreted unless the general pattern of the move-
acteristics of the equipment, the interpretation of ment is known. There is no direct relationship
the readings and the accuracy of the measurements between displacements of the landslide and the
are discussed as well. displacements of the wire. This is because the total
In the paper, we have followed the terminology displacement of the wire depends on several factors
for describing landslide types proposed by Dikau such as the diameter of the borehole, type of
et al. (1996). landslide mechanism, and the thickness, depth,
and dip angle of the surface of failure.
Displacements measured with the wire extensome-
2. Description of the borehole wire extensometer ter are global, and the device can neither detect
the vertical and the horizontal components of the
The wire extensometer is a simple and low-cost movement separately nor identify the presence of
device that allows the measurement of the relative several slip surfaces.
displacement between two points, one in the land- A theoretical analysis of the displacements of a
slide mass that is in motion and the other in stable wire for the case of a translational slide is presented
ground. The device that will be described here in Appendix A. The analysis has been developed
requires a borehole drilled through the landslide according to the following assumptions: (1) a
mass, up to the stable ground. It can be installed simple sliding mechanism occurs, consisting of the
together with either an inclinometer or a piezome- development of a single shear zone without or
ter casing. It consists of a protected steel wire with negligible internal deformation above it; (2)
anchored to a fixed point inside the borehole below a borehole, equipped with an extensometric wire,
the slip surface of the landslide. The opposite end has been drilled vertically through the landslide
of the wire is attached to a frame outside the body and crosses the surface of failure; (3) the
borehole. The frame is anchored to the landslide shear zone has a constant thickness and shows a
surface and holds a pulley. The wire is placed linear increase in the rate of displacement towards
around the pulley and is kept in tension by means the top. The rate of displacement is null at the
J. Corominas et al. / Engineering Geology 55 (2000) 149–166 151

Fig. 1. Sketch of the borehole wire extensometer device.

lower boundary of the shear zone and maximum lower end of the wire remains stable. The wire
at the upper boundary. length inside the hole is thus shortened, and the
Displacements of the wire may be either inside counterweight pulls the wire out of the bore-
or outside the borehole. The sense of displacement hole, causing the pulley rotate in a counter-
depends on the geometry of the surface of failure clockwise movement. As the landslide moves,
in relation to the wire. From now on, displace- the wire is deflected from its original position,
ments will be considered as positive when the wire and negative displacements will take place until
is introduced into the borehole, and as negative the angle between the wire and the vertical
when it is pulled out. equals the angle of inclination of the shear
We have given specific values to the equations surface, in other words, until the wire is located
of Appendix A for a landslide moving over a shear perpendicular to the slip surface.
zone dipping downwards in the direction of move- 2. In the second phase, wire deflection continues,
ment progression. In relation to the sense of dis- but now the wire is pulled into the borehole.
placement of the wire, two main phases are Positive increments will continue if the landslide
predicted (Fig. 2): geometry remains unchanged.
1. In the first phase, negative increments of wire During the first phase and at the beginning of
displacement occur. This is due to the down- the second phase, the displacements of the wire
ward component of the movement of the top are smaller than those measured at the ground
borehole-end and pulley support, while the surface and show a non-linear relationship. For
152 J. Corominas et al. / Engineering Geology 55 (2000) 149–166

Fig. 2. Relationship between the theoretical displacements of


the wire extensometer and the displacements of the ground sur-
face for a translational slide. The displacements of the wire have
been calculated using Eqs. (1)–(3) of Appendix A according to
the following parameters: H =10 m, H =500 mm b=10° and
m s Fig. 4. Relationship between the theoretical displacements of
w=60 mm.
the wire extensometer and the displacements of the ground sur-
face for a translational slide moving over a shear zone dipping
an angle ranging from −20° (upward ) to 20° (downward) in
landslide displacements much larger than the thick- the sense of movement. The rest of the parameters are the same
ness of the shear zone, the rates of displacement as those used in Fig. 2.
of both ground surface and wire tend to coincide,
although the overall (cumulative) displacement of
the latter will always be smaller ( Fig. 3). The hill are a particular instance of the case described
thicker the shear zone the greater are the differ- (H =0).
ences between both measurements. Translational s
Negative displacements of the wire may only
landslides with a single shear plane dipping down- occur when the movement develops over a down-
hill dipping shear zone. For a translational slide
moving over either a horizontal or upwards dip-
ping shear zone, only positive displacements are
predicted in the wire since the movements have
started (Fig. 4).

4. Monitoring with the wire extensometer at


Vallcebre and Alverà sites

The results of the performance of the wire


extensometer in the Vallcebre translational slide,
Eastern Pyrenees (Spain) and the Alverá mudslide,
Dolomites (Italy) will be presented below.

4.1. Geologic and geomorphologic context of the


Vallcebre landslide
Fig. 3. Relationship between the theoretical displacements of
the wire extensometer and the displacements of the ground sur-
face for a translational slide with a vertical component of the The landslide site is located 140 km north of
thickness of the shear zone (H ) ranging between 0 and 500 mm. Barcelona. The Vallcebre basin belongs to the
s
The rest of parameters are the same as those used in Fig. 2. imbricated thrust system of the southern Pyrenees
J. Corominas et al. / Engineering Geology 55 (2000) 149–166 153

(Muñoz et al., 1986) and, specifically, to the Lower both the ground surface and trees standing on it
Pedraforca thrust sheet that displays an east–west are tilted upslope. The toe of the landslide is being
trend. It consists mostly of sedimentary rocks of eroded by the Torrent of Vallcebre, originating a
Garumnian facies ( Upper Cretaceous–Lower minor scarp. There, the landslide foot reaches the
Palaeocene). torrent bed and overrides the opposite slope with
The Vallcebre landslide is of a translational the ground surface tilted backwards. The average
type. Fig. 5 shows a geomorphologic sketch of the slope of the whole landslide is about 10. The slide
landslide and the location of the monitored points mass is 1200 m long and 600 m wide.
and boreholes. It has a stair-shaped profile formed The direction of both the transverse scarps and
by four main morphological units of decreasing grabens is suggestive of a movement towards the
thickness towards the landslide toe. Each unit is northwest. A secondary direction of movement,
formed by a gentle slope surface bounded by a towards the Torrent Llarg, is also observed in the
secondary scarp of the landslide of a few tens of upper slide units. The most active area is the lower
metres high. At the toe of each scarp, there exists unit bounded, in the southwestern side, by the
an extension area that originates a graben in which torrents of Vallcebre and Llarg and, on its north-

Fig. 5. Geomorphologic sketch of the Vallcebre landslide.


154
J. Corominas et al. / Engineering Geology 55 (2000) 149–166

Fig. 6. Geological cross-sections of Vallcebre landslide interpreted from surface mapping and borehole logs.
J. Corominas et al. / Engineering Geology 55 (2000) 149–166 155

eastern side, by a well-developed lateral shear least 1.5–2 m within the limestone in order to
surface. ensure a proper anchorage of the wire. The wire
The geological structure of the slope in which was introduced into the borehole up to the bottom
the Vallcebre landslide is located has been obtained and then fixed to the ground with cement grouting.
either by mapping superficial exposures, from data To make sure that the wire reached the borehole
provided by geophysical surveys, or by interpreting end, a weight was attached to the wire. This kept
borehole logs (Fig. 6). The mobilized material the wire tensioned until it reached the borehole
consists of a set of shale, gypsum, and claystone bottom at the expected depth. The pulley system
layers gliding over a thick limestone bed. From of the wire extensometers was built at the work-
the bottom to the top, it includes: (1) densely shop of the Technical University of Catalonia
fissured shales, 1–6 m thick, showing slickensides ( UPC ) as a prototype based on designs made by
formed during the emplacement of the thrust fault- Angeli et al. (1988, 1989). The pulley incorporated
ing; (2) gypsum lenses up to 5 m thick and some a dial with marks every millimetre that would be
tens of metres long. These gypsum lenses are used to check the potentiometric readings. A com-
frequently 5–15 m above the sliding surface; and plete turn of the dial was equivalent to 400 mm
(3) clayey siltstones rich in veins and micronodules displacement of the wire. The potentiometer exhib-
of gypsum. In addition to these layers, in the ited a 1 kV electrical change for each 41 mm
extension zones located at the toe of the scarps, change in wire displacement. The accuracy
colluvium, composed of gravel with a silty matrix, achieved with the automatic reading (potentiome-
may be found. ter and data logger) was about 0.1 mm. The accu-
racy of the direct readings in the dial was about
4.2. Monitoring network of the Vallcebre landslide 0.5 mm. The upper end of the wire was placed
around the pulley and tensioned by attaching a
Vallcebre landslide has been monitored since sliding counterweight that was left hanging in a
1987 using conventional surveying and photogram- vertical hole 50 cm deep drilled on the other side
metry (Gili and Corominas, 1988). In July 1996, of the pulley system. When the counterweight was
a plan of drilling boreholes and installing a moni- close to the end of its travel — the pulley edge —
toring network was set out. The landslide was it could be relocated by sliding on the wire. In
equipped with 14 boreholes drilled into two sets order to measure large landslide displacements,
of seven boreholes. The first set was completed in 5 m of wire were left outside the borehole along
July 1996 and was arranged in an east–west direc- with the sliding counterweight. The potentiometric
tion close to the northern edge of the landslide. signal was recorded with a data logger system with
The second set was drilled in March–April 1997 adjustable time steps. Potentiometer readings were
and arranged parallel to the southern edge. taken automatically every 20 min and stored until
Borehole logs have confirmed the existence of the they were downloaded each 3–4 weeks.
‘grabens’ between the landslide units. For instance, Parallel to the installation of these instruments,
borehole S-7, located at the upper edge of the we carried out periodic monitoring of the Vallcebre
intermediate unit (Fig. 6), showed colluvium, 34 m landslide. Measurements were made with an incli-
in thickness, lying directly above the slip surface . nometer every 2–3 weeks until they went out of
The average inclination of the surface of failure of order and with a differential Global Positioning
landslide deduced from the results of inclinometer System (GPS ), at least once in 2 months (Gili
readings is 10° towards the Vallcebre torrent and et al., 1999, this volume). The latter included the
roughly parallel to the ground surface. measurement of displacements at the upper end of
Half of the boreholes were equipped with both the boreholes that had been used to analyse the
wire extensometers and open standpipe piezome- relationship between the superficial displacements
ters. It was assumed that the Vallcebre limestone, and the movement of the wire. Finally, rainfall
underneath the surface of failure, was stable and groundwater table changes were permanently
ground. Therefore, the boreholes were drilled at recorded as well.
156 J. Corominas et al. / Engineering Geology 55 (2000) 149–166

4.3. Performance of the wire extensometer in total accumulated (positive) displacement was less
Vallcebre than 10 mm during the observational period.
However, the actual horizontal displacements in
The extensometric wire device has proved to be the intermediate unit measured with a GPS were
very useful in recording sudden changes in rates significantly greater than those recorded by the
of displacements that can be directly related to the wire.
variations of the groundwater table and indirectly, The analysis of the displacements recorded at
to the rainfall records. For instance, sudden the boreholes S-2, S-5 and S-6 also shows that the
changes in the rate of displacement due to short wire has followed the two phases predicted for
rainy periods were recorded in their entirety. translational slides. At the beginning, boreholes
Displacements measured with the wire exten- S-5 and S-6 experienced negative increments of
someters at boreholes S-2, S-5 and S-6 are shown wire displacement (shortening of the wire length
in Fig. 7. The first two boreholes are located in inside the tube). Positive displacements were
the lower landslide unit, whereas the latter stands reached in a few weeks at S-5, although S-6 took
on the intermediate unit. During the observational several months. Borehole S-2 showed only positive
period (from November 1996 to December 1997), displacements. The behaviour of the latter is
the landslide never stopped. However, rates of explained by the timing of the completion of the
displacement were different at each morphological instrumentation works. The wires were anchored
unit and experienced significant changes in relation just after the completion of the boreholes in July
to the rainfall. The lower unit was the most active 1996, and the pulley systems and data loggers were
with an accumulated displacement of the wire of installed by the end of November of the same
over 500 mm in borehole S-2. The highest landslide
year. Because the landslide was active during this
activity period took place in the second half of
period of time, the first phase of negative wire
January and during February 1997, during which
displacement was already completed when the data
rates of up to 9 mm/day and 50 mm/week were
logger system was implemented at borehole S-2.
recorded in this borehole. The intermediate unit
Instead, the lower rate of displacement at S-6
was much less active. At the borehole S-6, the
allowed both phases to be recorded.
The relation between the groundwater level and
the rate of wire displacement is shown in Fig. 8.
There is a perfect synchronism between changes in
both records. However, from November 1996 to
January 1997, both rates are not proportional.
This is because it corresponds to the end of the
early stages of wire movement in which there is
no linearity between them and superficial displace-
ments (see Fig. 3). After January 1997, the rate of
displacement of the wire tends to be the same for
similar positions of the water table. The event of
June 1997 is an exception that apparently contra-
dicts this direct relationship. However, this is
because the episode of June 1997 was a very short-
lasting event. The complete groundwater table rise
Fig. 7. Wire extensometer measurements at boreholes S-2, S-5 and withdrawal lasted for only 14 h, whereas the
and S-6 of the Vallcebre slide versus time (each minor tick rate of wire displacement is given for a 24 h span.
corresponds to 1 week). Readings started 4 months after wire
extensometers were fixed to the borehole ends. Values below
In this case, the rate of displacement is smaller
zero observed at S-5 and S-6, correspond to the phase of nega- than the expected wire displacement if the ground-
tive displacements of the wire. water rise had lasted the whole day.
J. Corominas et al. / Engineering Geology 55 (2000) 149–166 157

4.4. Estimation of superficial landslide


displacements from the wire extensometer

Monitoring of superficial landslide displace-


ments was carried out by means of EDM and GPS
techniques. All the upper ends of the boreholes
were spatially referenced using a radiation method
with Real Time Kinematic GPS (Gili et al., 1999).
The results obtained were x, y (time) displacement
plots. The accuracy of the displacements measured
with GPS in the network implemented in Vallcebre
is estimated to be within 2 cm, which is similar (or
even better) to the accuracy provided by conven-
tional surveying using stable benchmarks located
on the opposite slope.
The measurements made with the GPS equip-
ment at the upper end of the boreholes were much
larger than those observed with the wire. For
instance, at borehole S-2 ( Table 1), a gap of about
15 cm occurred in only 2 months (from December
1996 to February 1997), increasing slightly until
June 1997. This gap is a consequence of the early
phase of negative wire displacements and is consis-
tent with the expected behaviour of the wire. After
June 1997, the observed increments with both
procedures tended to coincide.
Fig. 8. Above: groundwater table fluctuation at the borehole We have used the data recorded in the field to
S-2 versus time. Below: rate of wire displacement (solid line) check whether the Eqs. (1)–(3) of Appendix A
and of horizontal component of superficial displacement are able to reproduce the horizontal displacements
(dashed line). The latter has been calculated from the total wire of the landslide surface from readings of the wire
displacement using the equations of Appendix A. Each minor
tick on the time axis corresponds to 1 week.
at borehole S-2. All the data required for solving
the equations were known except for the thickness
of the surface of failure.

Table 1
Displacements at the borehole S-2 of the Vallcebre landslide measured with a Global Positioning System (GPS) and a borehole wire
extensometer

Date of GPS displacement GPS accumulated Wire extensometer Wire extensometer GPS:wire
measurement (mm) disp. (mm) displacement (mm) accumulated disp. (mm) ratio

07.24.96 First measurement 0.0 Installation of the wire


10.24.96 81.5 81.5
11.19.96 Beginning of recording 0.0
12.03.96 80.2 161.6 5.8 5.8
02.13.97 364.9 526.5 217.7 223.5 1.67
04.04.97 91.4 617.9 78.2 301.7 1.17
06.02.97 54.5 672.4 37.3 339.0 1.46
07.29.97 62.4 734.8 66.2 405.2 0.94
10.03.97 84.5 819.3 77.3 482.5 1.09
158 J. Corominas et al. / Engineering Geology 55 (2000) 149–166

Fig. 10. Plot of the temporal evolution of the horizontal compo-


nent of the displacements of the Valicebre landslide with the
Fig. 9. Fitting of the measured displacements (dots) to a calcu- GPS and the probe inclinometer (open circles and stars), and
lated curve relating both the horizontal and wire displacements the total wire displacements of the borehole S-2 (solid line).
for the borehole S-2 (see the text for explanation). The set of Each minor tick on the time axis corresponds to 1 week. A
curves (solid lines) has been prepared solving Eqs. (1), (2a), continuous record of the horizontal component of displace-
(2b) and (3) of Appendix A for several thicknesses of the shear ments has been calculated from the wire measurements using
surface. The geometrical parameters of the S-2 used in the equa- the Eqs. (8)–(10) of Appendix A (dashed line). Borehole
tions are: H =14.5 m; L =19.6 m; b=10°, w=58 mm; parameters are the same used in the case of Fig. 9.
m 0
X =30 mm and X =10 mm.
p b
the horizontal displacement of the landslide from
The equations were then solved considering the wire records. We have used these equations to
several thicknesses of the failure surface. The reproduce the movements at the surface of a
results have been expressed as a set of curves landslide at borehole S-2. Fig. 10 shows the epi-
relating the horizontal ground displacements at sodic measurements of the horizontal component
the landslide surface and the displacements of the of the displacements of the Vallcebre landslide
wire ( Fig. 9). The pairs of field measurements (measured with the GPS and probe inclinometer)
provided by the GPS and the wire extensometer and total wire displacements of the borehole S-2.
at borehole S-2 have been plotted and matched, A continuous record of the landslide horizontal
using a non-linear regression, to a curve corre- displacements (dashed line) has been calculated
sponding to a thickness of the surface of failure from the wire displacements using Eqs. (8)–(10)
of about 310 mm. of Appendix A. The validity of the calculated
The excellent fitting confirms the feasibility of curve is confirmed by its perfect fitting to the
using the equations of Appendix A for calculating displacements measured in the field.

Table 2
Comparison between topographical and extensometric measurements in Alverà landslide

Year Topographic survey GPS accumulated Wire extensometer Wire extensometer GPS:wire
displacement (cm) disp. (mm) displacement (mm) accumulated disp. (mm) ratio

1991 17.1 17.1 6.6 6.6 2.59


1992 14.9 32.0 6.9 13.5 2.16
1993 18.6 50.6 10.5 24.0 1.77
1994 14.7 65.3 10.1 34.1 1.46
1995 10.8 76.1 8.0 42.1 1.35
1996 9.4 85.5 7.5 49.6 1.25
J. Corominas et al. / Engineering Geology 55 (2000) 149–166 159

4.5. Wire extensometric measurements at the time spanning from the Carnian to the Rhaetian
Alverà mudslide (Pasuto et al., 1997). The alternance of ductile
and brittle rock types has developed morphological
The Alverà landslide affects a slope located near conditions favourable to the development of mass
Cortina d’Ampezzo in the Eastern Dolomites, Italy movements, some of them showing a recurrent
(Angeli et al., 1996a,b; Panizza et al. 1996). This activity since the retreat of the Würmian glaciers.
landslide is a rather elongated and active mudslide, For instance, a sample of peat collected at the slip
which, in the past, forced the adoption of emer- surface of the Alverà landslide was dated back to
gency measures for the village while the inhabitants 8710±70 years BP (Pasuto et al., 1997).
moved elsewhere (Fig. 11). The unstable slope is characterized by the pres-
The stratigraphical sequence cropping out in ence of clayey materials resulting from the weather-
the area of Cortina d’Ampezzo covers a period of ing of the S. Cassiano Formation, which mostly

Fig. 11. Topographic map of the Alverà landslide with the location of the boreholes.
160 J. Corominas et al. / Engineering Geology 55 (2000) 149–166

consists of alternating beds of sandstone, marls surface have revealed a high content of montmoril-
and clays directly outcropping in the upslope area. lonitic clay (Angeli et al., 1996a). Shear strength
The study of the area started in 1989 in tests performed on reconstructed samples gave w
r
co-operation with the Regione Veneto authorities values of 20–23°.
and was first directed to the protection of the
buildings located at the mudslide toe. At that time, 4.6. Effects of soil volume changes on the
several monitoring investigations and laboratory displacement of the wire at Alverà
tests were being carried out. The monitoring net-
work consisted of inclinometric casing and open The displacements in the boreholes equipped
(standpipe) piezometers equipped with electrical with extensometer devices were recorded every
transducers for the measurement of the hydraulic 10 min and stored in a data logger. The measure-
head in the slope. The inclinometers were provided ments obtained with the wire extensometer were
with a steel wire extensometers for the continuous checked with those obtained with both inclinome-
measurement of the landslide displacements. In ter and superficial surveying. The results showed
1992, a meteorological station was set up for significant differences that tend to stabilize with
recording precipitation (both rainfall and the time as expected after long displacements. Fig. 12
snow-water equivalent), air temperature and snow- shows both displacements recorded with the wire
cover thickness. In 1994, within the frame of and superficial surveying between 1990 and 1997
the TESLEC project funded by the European at the most superficial sliding surface located 5 m
Commission, ten boreholes were drilled in the deep. The average rate of displacement is slightly
mudslide to a depth between 9 and 30 m. The higher than 10 cm/year.
boreholes were arranged following the longitudinal Table 2 shows the annual displacements mea-
profile of the mudslide (seven open-pipe piezome- sured with both procedures and their ratio. A
ters) and in a cross direction (three inclinometers). progressive convergence of measured displacement
Each borehole was included in the network of is observed when comparing the ratio of the yearly
repeated topographic survey ( Fig. 11). readings of both systems.
The analysis of the borehole cores allowed the Some events of anomalous displacements have
identification of two separate layers and a main been observed as well. Fig. 13(d ) shows three
slip surface ranging from a depth of 18 to 25 m. episodes of positive wire displacement (wire
Another more superficial slip surface that appeared stretching) followed by negative wire displace-
to be very active, was identified at a depth of 5 m ments of the same order of magnitude few weeks
in the lower sector of the mudslide. The upper later. We have interpreted these displacements as
layer, about 20 m thick, in which the mudslide
occurred, consists of irregular, poorly-sorted
blocks of the original rock dispersed in an argilla-
ceous matrix and widely affected by cracks. The
lower layer consists of more consolidated homo-
geneous clays. A system of fissures filled with
calcite up to several centimetres thick is suggestive
of groundwater flow coming from the upper calcar-
eous slabs.
Soil mechanics laboratory tests were carried out
on samples taken from boreholes, drainage
trenches and trial pits. The results showed signifi-
cant differences between samples collected at any
depth in the slope and those taken from the slip Fig. 12. Alverà landslide, borehole S-5: comparison between
surface in a trial pit dug at the foot of the mudslide. extensiometric and topographic data recorded from 1990 to
Mineralogical analyses on samples form the slip 1997.
J. Corominas et al. / Engineering Geology 55 (2000) 149–166 161

a result of the effects of climate on the ground.


There is a seasonal coincidence — all of them
occurred in winter time — in these episodes.
Positive displacements of the wire were subsequent
to the groundwater table rise in the piezometer,
but there exists no relationship between them and
the preceding rainfall records [Fig. 13(a)]. Instead,
the positive displacements coincide with the occur-
rence of air temperatures below 0°C [Fig. 13(b)
(a) and (d )]. Therefore, all of these data suggest that
the cause of the positive increments of the wire is
the increase of the volume (swelling) of the upper
soil layer at the ground surface by freezing of the
water in the soil pores. Swelling leads to a rise of
both the ground surface and the pulley support,
with the subsequent displacement of the wire into
the borehole. Thawing of the interparticle ice
results in soil shrinkage and negative wire
increments.
In the Alverà mudslide, temperatures below 0°C
may last for several weeks, and the observed
(b) swelling of the ground surface may be as much as
1.5 cm. Freezing of the upper soil or rock layers is
not an unusual phenomenon in mountain regions.
For instance, in a rock outcrop in Montana,
Mikkelsen (1996) reported movements of the
ground by as much as 1.5 mm due to expansion
and contraction by seasonal changes in
temperature.

5. Conclusions
(c)
Borehole wire extensometer equipment is cheap
and simple. It enables very frequent readings that
may be stored in a data logger. The readings made
with the potentiometer allow an accuracy of
0.1 mm in the measurements.
Superficial landslide displacements may differ
significantly from those measured with the wire.
This is because the total displacement of the wire
depends on several factors such as the diameter of

(d) Fig. 13. Alverà landslide, borehole S-5: daily rainfall (a),
groundwater level (b), temperatures (c) and displacements (d)
recorded from 1994 to 1997. The influence of superficial frozen
ground on extensometric data is clearly observable.
162 J. Corominas et al. / Engineering Geology 55 (2000) 149–166

the borehole, the type of landslide mechanism, and Considering a downhill dipping surface of failure,
both the thickness and the dip angle of the surface two phases are expected: (1) negative displace-
of failure. ments of the wire are first produced due to the
A theoretical analysis of the relation between shortening of the borehole casing; and (2) positive
both wire and surgical landslide displacements for displacements take place when the angle between
translational slides has been carried out. the wire and the vertical line equals the inclination

Fig. 14. Stages of the position of the wire within the borehole casing in relation to the landslide displacements. The wire changes its
form when it meets a new contact with the borehole casing at the sides of the shear zone. Points M and F are located on the upper
and lower boundaries of the shear zone, respectively. Points B and P are fixed to the bottom of the borehole and to the pulley,
respectively. Each stage of the wire position is characterized by an specific equation.
J. Corominas et al. / Engineering Geology 55 (2000) 149–166 163

of the slip surface. This analysis also shows that 50 cm were recorded with the wire extensometer
when displacements are very large in relation to at borehole S-2 between November 1996 and
the thickness of the shear zone, the rates of dis- December 1997. All the wire extensometers
placement of the wire and at the landslide surface installed in Vallcebre were still functioning more
tend to coincide. than 1 year later.
The applicability of this analysis has been
checked in the field at the Vallcebre translational
landslide. Displacements of wire extensometers, Acknowledgements
when compared with measurements of superficial
displacements using differential GPS, follow the This work has been performed within the
relation predicted by the theoretical equations. A NEWTECH project, funded by the Commission
direct consequence of the latter is that both hori- of the European Communities, Environment and
zontal and vertical components of the landslide Climate Programme (Contract ENV-CT96-0248).
displacement can be estimated from the measure- The project joints the efforts from different groups
ment of the wire extensometer. ( France, Germany, Italy, Netherlands, Portugal,
The rates of displacement obtained with the Spain and United Kingdom) working in a few test
wire may be easily related to rainfall records and sites in Europe. The financial support from the
groundwater fluctuations. However, anomalous European Union is therefore gratefully acknowl-
wire measurements can be expected in case of edged. The additional support of the Spanish
freeze–thaw episodes of the upper soil layer as has Research Council (CICYT ) through contract
been shown in the Alverà mudslide. AMB96-2480-CE is gratefully acknowledged as
When combined with piezometers, the extenso- well.
metric wire has a great capability to reproduce the
response to the rainfall and groundwater fluctua-
tions with a high degree of accuracy. It also allows Appendix A: Analysis of wire displacement for a
temporal trends of the displacement to be observed theoretical translational slide
during critical rainy events.
The combined use of the wire extensometers Fig. 14 shows a theoretical sequence of wire
with devices such as probe inclinometers is highly displacement with the progression of a transla-
recommended. They complement each other very tional slide according to the following conditions:
well. Inclinometers have a short life when the (a) the slide moves as a rigid block over a single
landslide is very active, but they produce high- shear zone bounded by parallel planes, (b) the
quality information on landslide displacement pro- shear zone has a constant thickness and shows a
files, velocities and the position of the shear surface linear increase of the rate of displacement towards
immediately after its installation. Instead, at the the top. The rate of movement is null at the lower
early stages of deformation, the wire extensometer boundary of the shear zone and maximum at the
may only record negative displacements that are upper one, and (c) the wire is installed inside the
not directly related to the superficial ones. Once casing of a vertical borehole and fixed underneath
the inclinometer is lost, it allows a continuous the shear zone.
recording, and it is operational for very large The total displacement of the wire will be taken
displacements. The relationship found between the here as the change in length of the wire between
wire and superficial displacements enables con- points B and P (Fig. 14), corresponding to the
tinuous recording of both horizontal and vertical fixed end at the borehole bottom and to the first
displacements of the landslide to be achieved. point of contact with the pulley, respectively. In
The robustness of the equipment has been this analysis, the total displacement of the wire
proved at the Vallcebre landslide. While the incli- will be expressed as a function of the horizontal
nometer casings were out of order after movements component of the displacement observed at the
of less than 5 cm, total displacements of more than surface (point E by the borehole end).
164 J. Corominas et al. / Engineering Geology 55 (2000) 149–166

The length of the wire inside the borehole zone [point M of Fig. 14(c)], or at the lower
changes as the movement progresses. Different one (point F ). The wire will meet first point M
stages in changing the wire length may be iden- when the following condition is accomplished:
tified, depending on how the wire comes into
h ≤h tan h ≤tan h ,
contact with the borehole walls. Fig. 14 illustrates mp bf mp bf
the existence of three main evolutive stages. because
For a translational slide, with negligible defor-
mation above the shear zone, the superficial dis- X
tan h = p
placement, D , observed at point E is assumed to mp H +H
s m p
be equal to the displacement of the point M located
in the upper boundary of the shear zone. The path and
of D is parallel to the shear surface and, therefore, w−X
s b,
given the dip angle of the latter, it may be decom- tan h =
bf H
posed into a vertical and horizontal components f
(D and D ). Both components are related in the then
v h
following way:
X w−X
D =D tan b, p ≤ b,
v h H +H H
m p f
where b is the angle of dip of the shear zone (by
convention b will take a positive value when the rearranging the equation gives:

A B
landslide is moving downwards and a negative H
value in the opposite case). X f +X ≤w,
p H +H b
Changes in length of the wire inside the borehole m p
are dependent on the dimensions of the borehole where h and h are the angle measured from
casing (H , w), the distance of the wire ends to the mp bf
t the vertical of the lines linking respectively
casing (X , X ), and on the position, vertical com- points M and P, and B and F. Provided that
b p
ponent of the thickness and dip of the shear zone X , X ≥0 and ≤w. In case the condition above
(H +H , H , b). p b
m p s is not accomplished, the first contact with the
The changes in wire length during the different casing wall will be at point F.
stages are as follows: $ First contact at point M [Fig. 14(c)]:
1. Stage 1. Wire displacement prior to its contact The change in length of the wire (D ) between
with the casing [Fig. 14(b)]: l
points P and B is:
The change in length of the wire (D ) between
l D =L +L −L , (2a)
points P and B is: l mp bm o
D =L −L , (1) where
l bp o
where L ={(X2 +(H +H )2}1/2=cte.
mp p m p
L ={(D +X −X )2 and
bp h p b
+(H +H −D tan b)2}1/2 L ={(D −X )2
t p h bm h b
and +(H +H −(D −w) tan b)2}1/2.
f s h
L ={(X −X )2+(H +H )2}1/2. $ First contact at point F:
o p b t p
2. Stage 2. Wire displacement after its contact D =L +L −L , (2b)
l bf fp o
with the casing at one single point:
where
There are two possible cases: the contact is
established at the upper boundary of the shear L ={(w−X )2+H2 }1/2=cte
bf b f
J. Corominas et al. / Engineering Geology 55 (2000) 149–166 165

and then

L ={(D +X −w)2+(H +H +H X −X +d X
fp h p m p s p b 1,2a = p , (5)
−(D −w) tan b)2}1/2. H +H −d tan b H +H
h t p 1,2a m p
L and L are the distances between points rearranging and solving Eq. (4) for d :
bf fp 1,2a
B and F, and points F and P, respectively.
X (H −H )+X (H +H )
3. Stage 3. Wire displacement after its contact d = p t m b m p . (6)
with the casing at points M and F [Fig. 14(d )]: 1,2a H +H +X tan b
m p p
D =L +L +L −L , (3) Similarly, the boundary between stages 2a and 3
l mp s bf o is given by the following conditions:
where
H (w−X )
h =h d =w+ s b . (7)
L ={(D −w)2+(H −(D −w) tan b)2}1/2. bm bf 2a,3 H +(w−X ) tan b
s h s h f b
Eqs. (1), (2a), (2b) and (3), expressing the In case of a translational slide moving over a shear
change in length of the wire (D ) between points plane (H =0), the expression is reduced to:
l s
P and B, can be written in a general way as
follows: d #w.
2a,3
D ={(D +a)2+(b−D tan b)2}1/2+c, (4) However, the boundary between stage 1 and 2b
l h h
may be obtained from:
where a, b and c are geometrical constants that
are different for each displacement stage. (w−X )(H +H )−(X −X )H
h =h d = b t p p b f.
bp bf 1,2b H +(w−X ) tan b
f b
A.1. Range of applicability of the equations
Similarly, the boundary between stage 2b and 3 is
derived from:
The range of applicability of the Eqs. (1), (2a),
(2b) and (3) is defined by the geometrical charac- (H +H )w+(H +w tan b)X
teristics of the borehole and shear surface. h =h d = m p s p.
fp mp 2b,3 H +H +X tan b
The boundary between stages 1 and 2a is pro- m p p
duced when h =h . Being h and h , the angles In case of a translational slide moving over a shear
bp mp bp mp
that the deflected wire segments BP and MP form plane (H =0), the expression is reduced to:
respectively with the vertical. The horizontal dis- s
placement, d , that corresponds to the boundary d #w.
1,2a 2b,3
is:

h =h tan h =tan h A.2. Calculation of displacements at the ground


bp mp bp mp surface from the wire extensometer
because
Displacements of point E [Fig. 14(a)] can be
X −X +d calculated by rearranging the Eqs. (1), (2a), (2b)
tan h = p b 1,2a
bp H +H −d tan b and (3). The horizontal component (D ) of the
h
t p 1,2a landslide displacement at the ground surface is
and given by:

X D ={b tan b−a±


tan h = p , h
mp H +H 앀(a−b tan b)2−(a2+b2−(D −c)2)/cos2 b} cos2 b,
m p l
166 J. Corominas et al. / Engineering Geology 55 (2000) 149–166

where Angeli, M.G., Gasparetto, P., Menotti, R.M., Pasuto, A.,


Silvano, S., 1996a. A visco-plastic model for slope analysis
a=X −X , b=H +H , and c=−L for stage 1 applied to a mudslide in Cortina d’Ampezzo, Italy. Quart.
p b t p 0 J. Eng. Geol. 29, 233–240.
(8) Angeli, M.G., Gasparetto, P., Menotti, R.M., Pasuto, A.,
Silvano, S., 1996b. Examples of mudslides on low-gradient
a=−X , b=H +H +w tan b, and clayey slopes, Senneset, K. ( Ed.), Landslides. Proc. 7th Int.
b f s Symp. Landslides, Trondheim Vol. 1. A.A. Balkema, Rot-
c=L −L for stage 2a (9a) terdam, pp. 141–145.
mp 0
Bonnard, C., Steinmann, G., 1990. Continuous measurement
a=−X −w, b=H +H +H +w tan b, and of landslide movements, Geotechnical Instrumentation in
p p m s
Practice. Institution of Civil Engineers. Thomas Telford,
c=L −L for stage 2b (9b)
bf 0 London, pp. 177–189.
Dikau, R., Brunsden, D., Schrott, L., Ibsen, M.L., 1996. Land-
a=−w, b=H +w tan b, and
s slide Recognition. Identification, Movement and Causes.
c=L +L −L for stage 3 (10) Wiley, Chichester, UK.
mp bf 0 Gili, J.A., Corominas, J., 1988. Aplicación de técnicas fotogra-
The above equations show that two values of D métricas y topográficas en la auscultación de algunos desli-
h zamientos, III Simposio Nacional sobre Taludes y Laderas
are obtained for every value of D . This can be
l Inestables. La Coruña Vol. 3, 941–952.
seen in Fig. 2, in which one value of D corres- Gili, J.A., Corominas, J., Rius, J., 1999. Using Global Position-
h
ponds to the negative phase of wire displacement ing System techniques in landslide monitoring. Eng. Geol.
and the other to the positive phase. 55, 167–192. (this issue).
Mikkelsen, P.E., 1996. Field instrumentation. In: Turner, A.K.,
Schuster, R.L. ( Eds.), Landslides. Investigation and mitiga-
tion. TRB Special Report 247. National Academy of Sci-
ences, Washington DC, pp. 278–316.
References Muñoz, J.A., 1986. Thrust sequences in the Spanish eastern
Pyrenees. J. Struc. Geol. 8 3–4, 399–405.
Angeli, M.G., Gasparetto, P., Silvano, S., Tonetti, G., 1988. Panizza, M., Pasuto, A., Silvano, S., Soldati, M., 1996. Tempo-
An automatic recording system to detect the critical stability ral occurrence and activity of landslides in the area of Cor-
of slopes, Proc. 5th Int. Symposium on Landslides, Lau- tina d’Ampezzo (Dolomites, Italy). Geomorphology 15 3/
sanne Vol. 1. A.A. Balkema, Rotterdam, pp. 375–378. 4, 311–326.
Angeli, M.G., Gasparetto, P., Pasuto, A., Silvano, S., 1989. Pasuto, A., Siorpaes, C., Soldati, M., 1997. I fenomeni franosi
Examples of landslide instrumentation, Proc. 12th Int. Conf. nel quadro geologico e geomorfologico della conca di Cor-
Soil Mechanics and Foundation Engineering. Rio de Janeiro tina d’Ampezzo (Dolomiti, Italia). Il Quaternario 10 (1),
Vol. 3. A.A. Balkema, Rotterdam, pp. 1531–1534. 75–92.

View publication stats

You might also like