geosciences

Article
Using Electrical Resistivity Tomography Method to Determine
the Inner 3D Geometry and the Main Runoff Directions of the
Large Active Landslide of Pie de Cuesta in the Vítor
Valley (Peru)
Yasmine Huayllazo 1, * , Rosmery Infa 1 , Jorge Soto 1 , Krover Lazarte 1 , Joseph Huanca 1 , Yovana Alvarez 1
and Teresa Teixidó 2

1 Universidad Nacional de San Agustín de Arequipa, Arequipa 04000, Peru; [email protected] (R.I.);
[email protected] (J.S.); [email protected] (K.L.); [email protected] (J.H.);
[email protected] (Y.A.)
2 Andalusian Institute of Geophysics of Granada University, 18071 Granada, Spain; [email protected]
* Correspondence: [email protected]

Abstract: Pie de Cuesta is a large landslide with a planar area of 1 km2 located in the Vítor district,
in the Arequipa department (Peru), and constitutes an active phenomenon. It belongs to the ro-
tational/translational type, which concerns cases that are very susceptible to reactivation because
any change in the water content or removal of the lower part can lead to a new instability. In this
context, a previous geological study has been decisive in recognizing the lithologies present and
understanding their behavior when they are saturated. But it is also necessary to know the inner
“landslide geometry” in order to gusset a geotechnical diagnosis. The present study shows how the
deep electrical profiles (ERT, electrical resistivity tomography method), supported by two Vp seismic
refraction tomography lines (SVP), have been used to create a 3D cognitive model that would allow
Citation: Huayllazo, Y.; Infa, R.; Soto,
the identification of the inner landslide structure: the 3D rupture surface, the volume of the sliding
J.; Lazarte, K.; Huanca, J.; Alvarez, Y.;
mass infiltration sectors or fractures, and the preferred runoff directions. Moreover, on large landsides,
Teixidó, T. Using Electrical Resistivity
Tomography Method to Determine
placing the geophysical profiles is a crucial aspect because it greatly depends on the accessibility of
the Inner 3D Geometry and the Main the area and the availability of the physical space required. In our case, we need to extend profiles
Runoff Directions of the Large Active up to 1100 m long in order to obtain data at greater depths since this landslide is approximately
Landslide of Pie de Cuesta in the 200 m tall. Based on the geophysical results and geologic information, the 3D final model of the
Vítor Valley (Peru). Geosciences 2023, inner structure of this landslide is presented. Additionally, the main runoff water directions and the
13, 342. https://doi.org/10.3390/ volume of 90.5 Hm3 of the sliding mass are also estimated.
geosciences13110342

Academic Editors: Haijia Wen,

Keywords: large landslide; electrical resistivity tomography (2D-ERT); Vp-seismic refraction
Weile Li, Chong Xu and tomography (2D-SVP)
Jesus Martinez-Frias

Received: 26 September 2023

Revised: 30 October 2023
1. Introduction
Accepted: 3 November 2023
Published: 9 November 2023
Landslides are geological phenomena that consist of downhill earth movements
influenced by gravity and caused by natural and anthropogenic factors such as rainfall
episodes, agricultural activities, and civil land activities, among others [1–3]. These complex
phenomena have a large socioeconomic impact and require multidisciplinary geosciences
Copyright: © 2023 by the authors. research to determine their internal structure and surrounding environment to facilitate
Licensee MDPI, Basel, Switzerland. stability analysis and risk mitigation [4–6].
This article is an open access article From a geological standpoint, landslide occurrence is frequently related to shear
distributed under the terms and
strength reduction and fluidization processes suffered by clayey soils when water flux
conditions of the Creative Commons
occurs through the subsurface [7–11]. To understand the occurrence and evolution of a
Attribution (CC BY) license (https://
landslide, it is of prime importance to obtain information about the local geology, subsurface
creativecommons.org/licenses/by/
hydrogeological conditions, and depth of the failure surface. Frequently, such information
4.0/).

Geosciences 2023, 13, 342. https://doi.org/10.3390/geosciences13110342 https://www.mdpi.com/journal/geosciences

Geosciences 2023, 13, 342 2 of 19

cannot be easily obtained because it is needed to drill expensive boreholes over the landslide
zone. In this context, geophysical methods are increasingly being used with great success
to define the physical properties of soils [12,13] and to obtain geologic and hydrogeological
information about the study area [14,15]. Nowadays, landslides are also monitored in real
time using different remote sensors to predict their evolution [16,17].
The use of geophysical methods to study landslides presents several advantages:
(i) they are non-destructive techniques, (ii) they can reach the maximum target depth of
interest as an extensive mode, and (iii) the resulting models may show the inner geometry
of landslide mass, even in cases of structural complexity. Among the set of prospecting
methods available, the 2D Electrical Resistivity Method (2D-ERT) is very useful in the study
of landslides because it is sensitive to variations in water content and the grain size of
materials, providing detailed subsurface models [18,19]. Meanwhile, Seismic Vp Velocity
(2D-SVP) is a useful surface method for obtaining information on the different compactions
of materials to establish stability studies of the sliding mass. And in cases such as this,
where the flow water is salinized, this allows discrimination at the top of the basal layer [19].
However, this method is used less in landslide studies [20–22].
Peru, due to its geographical location and geomorphological conditions, is a country
that is exposed to several natural phenomena, where mass movements are one of the most
common events that cause human and economic damage. One of the largest and most
destructive landslides in the Vítor Valley (Arequipa) is the Pie de la Cuesta landslide, which
currently has a main scarp that is 1 km long and a planar area of approximately 1 km2 .
This mass movement has been taking place since 6 October 1974 [23], having produced a
large-scale collapse on 9 January 1975 that buried the town of Pie de Cuesta and affected
more than 12 families who had to be relocated [24]. In 2018, the Perú National Research
Program funded the National University of San Agustín de Arequipa to carry out a research
project (IBAIB-03-2018-UNSA) to characterize this landslide.
For this purpose, a geophysical survey consisting of electrical and seismic profiles was
carried out on the landslide to obtain the main inner geometry of the sliding layers with
these non-destructive methods. This paper summarizes the study carried out to correlate
electrical (2D-ERT) and seismic (2D-SVP) profiles with geological information and to obtain
the internal structure of a large landslide and its main water flow directions.

2. Study Area
2.1. Geological Description of Study Area
The Pie de Cuesta landslide is located at the foothills of agricultural terrain corre-
sponding to the La Joya Antigua system (Figure 1), which is a plain of more than 3500 ha of
crops where the irrigation method is mostly flooding. So, the accumulation of groundwater
is caused by irrigation. Geologically [24], the study area is formed by three formations
(Fm) (Figure 1): At the bottom, the Lower Moquegua Fm (96 m thick) is mainly composed
of polymathic conglomerates of rounded and surrounded clasts with a sandy matrix and
intercalated with arkosic sandstones and shales. In the middle, the Upper Moquegua Fm
(100 m thick) is placed. This formation is composed of red shales alternated with gypsum
layers, whose upper part contains red shales with layers of thick tuffaceous sandstones.
The top formation is the Millo Fm (55 m thick), which is mainly composed of conglomerates
of rounded clasts of intrusive and volcanic origin with a sandy matrix intercalated with
silty sandy gravels; a layer of cream-colored ignimbrite is also observed. In addition to
these formations, the study area also presents alluvial, coluvial and fluvial materials.
 In the main wall of the landslide, it can be seen that the water seeps into the Upper
Moquegua Fm through fractures and cracks. Two important water tables can also be
distinguished: the first one is in the Millo Fm and Lower Moquegua Fm contact (1492 m
Geosciences 2023, 13, 342 3 of 19
s.l.), and the second is in the Upper and Lower Moquegua Fms contact (1336 m s.l.), re‐
sulting in an interface of high humidity.

Figure 1.
Figure 1. Geological formations of
Geological formations of the
the Pie
Pie de
de Cuesta
Cuesta landslide.
landslide. On
On the
the left,
left, the
the stratigraphic column
stratigraphic column
has a 250 m thickness. On the right, the geological outcrops over the ortofoto‐map. In the northwest
has a 250 m thickness. On the right, the geological outcrops over the ortofoto-map. In the northwest
are the irrigated crops of the Joya Antigua, which act as a source of landslide activation.
are the irrigated crops of the Joya Antigua, which act as a source of landslide activation.

If wecroplands
The consider are
the placed
previous geological
over the Millodescription,
Fm, whichthis meansasthat
behaves a 250 m thick
a permeable unit,
column of geological materials is involved in this landslide. This constraint implies that
allowing the filtration of irrigation water due to its composition and poor consolidation. In
the main
the lengths ofofthe
wall thegeophysical
landslide, it profiles must
can be seen bethe
that long enough
water seepsto guarantee
into the Upper a subsurface
Moquegua
depth
Fm of inspection
through fracturesbelow the second
and cracks. water table:
Two important about
water 200 can
tables m inalso
depth.
be distinguished: the
first one is in the Millo Fm and Lower Moquegua Fm contact (1492 m s.l.), and the second
2.2.
is inDescription of theLower
the Upper and Landslide through Fms
Moquegua Historical
contact Aerial
(1336Imagery
m s.l.), resulting in an interface of
high The
humidity.
Cuesta landslide began in 1975 and remained active until the 1990s [25]. In 2016,
If we consider
it reactivated, whichthe hasprevious
continued geological
up to thedescription,
date of this thisstudymeans that a 250
[26]. Figure m thick
2 shows the
column
historicalof geological
evolution materials
of this is involved in
landslide. The thisfirst
landslide.
imageThis constraint implies
corresponds to a 1945 that the
or‐
lengths of the geophysical
tho‐photomap showing the profiles
state ofmust
the be long before
terrain enoughthe to guarantee
catastrophic a subsurface depth of
failure occurred in
inspection
1975 [27,28].below the second
The second imagewater table:a 2015
is from aboutGoogle
200 m in depth.
Earth photo, prior to the 2016 reac‐
tivation. And the third image is an orthomosaic from 2019, obtained from photogram‐
2.2. Description
metric of the
processing of Landslide throughInHistorical
a drone flight. the first Aerial
map ofImagery
1945, the landslide boundaries of
The Cuesta landslide began in 1975 and
the years 2015 and 2019 have been marked, showing the remained active until the 1990s
characteristics [25].landslide
of the In 2016,
it reactivated,
advance: which
between hasand
1975 continued
2015, theup to the
major date ofoccurs,
collapse this study [26].the
sliding Figure 2 shows
materials the
toward
historical evolution
the SE margin of direction
in the this landslide.
of theThe first image
valley, while thecorresponds
reactivationto a preference
1945 ortho-photomap
is for 2019
showing the state
and especially of the
of the terrain
main lobebefore
that isthe catastrophic
close to the rightfailure
flank occurred in 1975 [27,28]. The
of the landslide.
second image is from a 2015 Google Earth photo, prior to the 2016 reactivation. And the
third image is an orthomosaic from 2019, obtained from photogrammetric processing of a
drone flight. In the first map of 1945, the landslide boundaries of the years 2015 and 2019
have been marked, showing the characteristics of the landslide advance: between 1975 and
2015, the major collapse occurs, sliding the materials toward the SE margin in the direction
of the valley, while the reactivation preference is for 2019 and especially of the main lobe
that is close to the right flank of the landslide.
Geosciences
Geosciences2023,
2023,13,
13,x342
FOR PEER REVIEW 4 4ofof19
19

(a) (b) (c)

Figure
Figure2.2.Pie
PiededelalaCuesta
Cuestalandslide historical
landslide evolution.
historical (a) Ortho‐photomap
evolution. (a) Ortho-photomap of 1945 showing
of 1945 the
showing
primitive slope limits (black line), the landslide boundaries relative to the first stage at 2015 (orange
the primitive slope limits (black line), the landslide boundaries relative to the first stage at 2015
line), and the limits of the last reactivation in 2019 (yellow line). (b) Google Earth image corre‐
(orange line), and the limits of the last reactivation in 2019 (yellow line). (b) Google Earth image
sponding to 2015. (c) Orthomosaic of 2019, obtained from aerial photogrammetry using a drone.
corresponding to 2015. (c) Orthomosaic of 2019, obtained from aerial photogrammetry using a drone.
3.3.Materials
Materialsand andMethods
Methods
Figure 3 shows
Figure 3 shows thethe locations
locationsof ofgeophysical
geophysicalsurveys
surveysconducted
conducted inin this
this study,
study, where
where the
the
mainmain method
method applied
applied waswasthe the electrical
electrical resistivity
resistivity tomography
tomography profiles
profiles (2D‐ERT),
(2D-ERT), as
as they
they provide
provide suitable
suitable electrical
electrical resistivity
resistivity models
models that that are sensitive
are sensitive to lithological
to lithological changes
changes and
and the water
the water content
content of theofmaterials.
the materials.
And itAnd it is useful
is useful in complex
in complex geological
geological conditions
conditions [29,30].
[29,30]. As a complementary method, two velocity P‐wave seismic refraction
As a complementary method, two velocity P-wave seismic refraction tomography surveys tomography
surveys
(2D-SVP) (2D‐SVP) wereout
were carried carried out in
in order order to evaluate
to evaluate the safetythe safety
of the main of materials
the main present
materialsin
present in the [31,32].
the landslide landslide [31,32].

3.1. Geophysical Non-Invasive Data Acquisition

The purpose of electrical surveys is to determine the subsurface resistivity distribution
by making measurements on the ground surface. The electrical c.c. surveys were acquired
by a set of electrodes placed on the ground at defined distances, as shown in Table 1. In
general terms, the method consists of injecting a current into the ground using two current
electrodes and measuring the resulting voltage difference of the generated electric field at
two other electrode pairs of the section (potential electrodes). With the Ohm law and the
inversion procedure, the result is a parametric 2D model with ground-apparent resistivity
distributions. As seen in Figure 3c, the electrical dataset was acquired using an Elect
Pro-10 resistivimeter (Iris Instruments, Inc.; Orleans, France) which is a 10-channel
(b) receiver
especially designed to record deep profiles. In this study, the resistivimeter was used in
conjunction with the Elec Pro Switch, which allows 48 electrodes to be connected to the
device by multi-electrode cables. The receiver pulse signal was 2 s, and the maximum input
voltage was 15 V (automatic gain). For the measurements, we used a pole-dipole electrode
array configuration, which has good horizontal coverage, although it has a higher signal
compared to others electrode configurations. It is suitable for this deep work because it is
not sensitive to telluric noise [33]. The main technical features of this data acquisition are
summarized in Table 1.

(a) (c)

Figure 3. (a) Locations of geophysical surveys in the study zone. The 2D‐ERT profiles are marked
with yellow lines, and the 2D‐SVP sections are marked with blue lines. Note that the 2D‐SVP
 the main method applied was the electrical resistivity tomography profiles (2D-ERT), as
they provide suitable electrical resistivity models that are sensitive to lithological changes
and the water content of the materials. And it is useful in complex geological conditions
[29,30]. As a complementary method, two velocity P-wave seismic refraction tomography
Geosciences 2023, 13, 342
surveys (2D-SVP) were carried out in order to evaluate the safety of the main5 materials
of 19

present in the landslide [31,32].

(b)

(a) (c)

Figure 3. 3.(a)(a)Locations
Figure Locationsofof geophysical surveysininthethe
geophysical surveys study
study zone.
zone. TheThe 2D-ERT
2D-ERT profiles
profiles are marked
are marked
with yellow
with yellowlines, andthe
lines, and the 2D-SVP
2D-SVP sections
sections are marked
are marked with
with blue lines.blue
Notelines. Note
that the thatmatches
2D-SVP the 2D-SVP
the 2D-ERT-01 and 3D-ERT-03 profiles. (b) The absence of sections in the southeast sector of the
landslide is due to the rugged topography; see image. (c) The electrical equipment is especially
designed to reach the required depth, mainly consisting of an Elect Pro-10 resistivimeter (blue box in
the foreground), a channel switch-Pro (orange box in the background), and a current injector device
(at the back).-Iris Instruments, Inc. (2023) from www.iris-instruments.com/elrec-pro.html (accessed
on 29 October 2023).

Table 1. Main technical features of the geophysical equipment and data acquisition.

Reached
Profiles Sensors Spacing # of Sensors Sensor Array Total Length
Depth
2D-ERT01 Electrodes at 40 m 33 Pole-Dipole 640 m 160 m
2D-ERT02 Electrodes at 30 m 35 Pole-Dipole 510 m 120 m
2D-ERT03 Electrodes at 25 m 45 Pole-Dipole 550 m 140 m
ERT survey 2D-ERT04 Electrodes at 25 m 29 Pole-Dipole 350 m 90 m
2D-ERT05 Electrodes at 30 m 27 Pole-Dipole 390 m 100 m
2D-ERT06 Electrodes at 50 m 45 Pole-Dipole 1100 m 220 m
2D-ERT07 Electrodes at 50 m 35 Pole-Dipole 850 m 200 m
2D-SVP01 Geophone at 5 m 2 × 24 of 20 Hz 13 shots 420 m 240 m
SVP survey
2D-SVP02 Geophone at 5 m 2 × 24 of 20 Hz 19 shots 600 m 70 m

For the seismic method, the Vp refraction tomography considers the first arrival times
of the P-waves produced by a controlled source at different sensor stations (geophones).
In this study, the seismic 2D-SRT profile was acquired with a Geode system (Geometrics
Inc.; San Jose, USA) that controls 24 vertical geophones of 20 Hz of natural frequency. We
used custom-made low-energy explosives as the seismic source in order to achieve a good
Geosciences 2023, 13, 342 6 of 19

signal at far offsets. The result was a 2D seismic model of the subsurface showing the Vp
velocity distribution [34].
The spread of the seismic recording data (Table 1) was designed in order to ensure ray
coverage along the entire section. Accordingly, the center shot positions were repeated to
link the recording units with 24 channels each. The sample rate was 0.125 ms, and the total
recording time was 0.5 s. The seismic sources were placed at 0.5 m depth.

3.2. Electrical Data Processing

The ERT method calculates the electrical resistivity properties of rocks beneath the
surface and is a well-established method in near-surface characterization studies [34]. In
this study, to calculate the resistivity models from the apparent field-measured parameters,
the Res2dinv commercial software (V5.0, Bentley Systems; Exton, PA, USA) was used.
This code is designed to obtain 2D (and 3D) geoelectric models by applying the inversion
computational technique [35,36]. Due to the fact that the subsurface resistivity is strongly
influenced by a rock’s properties such as porosity, mineral composition, fluid content,
and fault structure [37,38], different processing options have been tested for the inversion
procedure. In particular, we wished to determine the influence of the finite mesh grid size
and the effect of the damping inversion factors. The damping factor leads to a stabilization
of the solution but produces a smoothed resistivity model [37]. In our case, a smoothed
section was not desirable because a large landslide is characterized by a heterogeneous
underground. Consequently, we selected medium damping factors to stabilize the calcula-
tion with a maximum mesh refinement of the parameter models. Other aspects included
the choice of the inversion algorithm. When the subsurface has vertical discontinuities,
such as falls, the conventional least squares smoothness-constrain method tends to smear
the boundaries, and it is best to operate with the robust constrained inversion method,
which is less sensitive to resistivity contrasts but gives a high apparent resistivity, although
with more fitting errors [38]. Finally, we opted for the following inversion parameters: a
refinement cell model defined by four nodes for electrode spacing, a medium inversion
damping factor (0.15), and a robust constrained method. The convergence iterations and the
absolute error between the measured and calculated apparent resistivities are summarized
in Table 2. Figure 4 is an example of the inversion processing for the 2D-ERT-IP-02 profile.

Table 2. Inversion parameters of the geophysical models.

2D -Sections # of Iterations Abs. Error Reached Depth (m)

2D-ERT01 8 8.9 167
2D-ERT02 7 8.6 120
2D-ERT03 9 6 125
ERT survey 2D-ERT04 10 8.8 95
2D-ERT05 9 9.8 126
2D-ERT06 10 12.3 215
2D-ERT07 10 19.8 210

In this study, the 2D-ERT-IP-06 and ERT-IP-07 electrical profiles presented more errors;
paradoxically, these two surveys are located outside the landslide behind the northern scar,
where a horizontal stratification is assumed (Moquegua Fm). As it will be seen, the cause is
due to a fault system detected in this setting.
 strained inversion method, which is less sensitive to resistivity contrasts but gives a high
apparent resistivity, although with more fitting errors [38]. Finally, we opted for the fol‐
lowing inversion parameters: a refinement cell model defined by four nodes for electrode
spacing, a medium inversion damping factor (0.15), and a robust constrained method.
The convergence iterations and the absolute error between the measured and calculated
Geosciences 2023, 13, 342 7 of 19
apparent resistivities are summarized in Table 2. Figure 4 is an example of the inversion
processing for the 2D‐ERT‐IP‐02 profile.

Figure 4. General information resulting from the inversion processing flow for the 2D‐ERT‐02 pro‐
Figure 4. General information resulting from the inversion processing flow for the 2D-ERT-02 profile.
file. The histogram shows the absolute error (%) between calculated and measured apparent resis‐
The histogram shows the absolute error (%) between calculated and measured apparent resistivities
tivities in the 7th iteration. The upper text (“Inversion results”) indicates the convergence in suc‐
in the 7th
cessive iteration.
iterations. The
The uppergraph
second text (“Inversion results”)
is the correlation plotindicates
betweenthe convergence
measured in successive
and calculated ap‐
iterations.
parent The second
resistivities in thegraph is the correlation plot between measured and calculated apparent
7th iteration.
resistivities in the 7th iteration.

3.3. Seismic Refraction Data Processing

P-wave travel time tomography is currently a well-established and broadly used inver-
sion scheme to resolve Vp velocity structure [39]. Travel time tomography is a non-linear
problem in geophysics because the deflection of seismic rays depends on the unknown
velocity of the subsurface structure. Common approaches use inversion algorithms con-
sisting of picking the first arrival travel times of the P-waves and searching for the most
plausible velocity model that can reproduce the observables by minimizing the time differ-
ence between the estimated travel times. Theoretical travel times are thus calculated using
a ray-tracing forward modeling scheme.
In our study, the first arrivals were handpicked from the shot records (Figure 5a), and
we used a commercial Rayfract code (Intelligent Resources Inc. Software, Winnersh, UK)
for the inversion proposal (Table 3). To determine the subsurface velocity distribution from
the first arrivals picked travel times, we used the Delta-t-v inversion method [40]. This
technique is based on the common mid-point (CMP) refraction concept, which considers
the CMP travel times as a function of the independent variable CMPx coordinates and the
CMP constant offsets (Figure 5c). It starts by determining the velocity at the base of a layer
from CMP travel time curves (Figure 5b), and then it numerically inverts the velocity at the
top of the gradient layer. The algorithm automatically identifies precise time delays on CMP
curves, transforming these delays into velocity–depth anomalies. A 1-D Vp velocity–depth
function is constrained beneath each CMP. All 1D velocity–depth functions are integrated
through a gridding scheme, building up a final 1.5D velocity model. A simple and smooth
1D velocity model is needed to initialize the process. This is obtained by laterally extending
a simple 1D layered model along the profile.
Geosciences 2023, 13, x FOR PEER REVIEW 8 of 19
Geosciences 2023, 13, 342 8 of 19

(a)

(b) (c)
Figure 5. Basic inputs to resolve Vp subsurface distribution. (a) Examples of the first time picked
Figure 5. Basic inputs to resolve Vp subsurface distribution. (a) Examples of the first time picked
for the P‐waves in two shots viewer (red crosses). (b) Space–time graphs of all picked first arrivals.
for the P-waves in two shots viewer (red crosses). (b) Space–time graphs of all picked first arrivals.
(c) The first time picks as a function of the independent variables CMPx (station) and the CMP
(c) The first
constant time
offset; picks
curves areaspresented
a function of the
with independent
a velocity variables
reduction of 4500 CMPx
m/s. (station) and the CMP
constant offset; curves are presented with a velocity reduction of 4500 m/s.
4. Results and Interpretation
Table 3. Seismic fitting between modeled and picked travel times. Normalized RMS error is the RMS
Landslides are considered a natural process that occur in a variety of geologic set‐
error, divided by average pick time of all traces modeled.
tings, either as soil mass movement, debris flow, rockfall, or combinations of both. In this
case, the main triggering
Profile factor has
Normalized RMSbeen the continuous
Error agricultural irrigation
# Traces Modeled of the crop
# of Iterations
fields located to the north (La Joya Antigua, Figure 1). So, for years (around 1968 until
2D-SVP01 8.5 (%) 279 8
P-wave seismic survey today), irrigation water has been infiltrating the ground through surface cracks and sink
2D-SVP02 9.1 (%) 281 7
holes to saturate the subsurface. In fact, irrigation not only triggers landslides in this area,
but the occurrence of a landslide itself also increases the chances for new events to hap‐
pen, asFigure 5c shows
observed in thethe first arrival
reactivation of times in the CMP
2016 (Figure 2). scheme. In this way, the effects of
dipping layers are averaged and minimized. The travel times are smoothed by stacking
CMP-sorted
4.1. travel
Basis for the time curves over 40 adjacent CMPs. Subsequently, each curve is “Deltat-
Interpretation
v inverted”.
Landslides Table
are3classified
summarizes the quantitative
by their indicators
type of movement in the
[41,42], Vpactually
and final models, which
the scheme
suggests that they have a relatively high degree of reliability.
terminology is suggested by the UNESCO Working Party on the ‘World Landslide In‐
ventory’ (WP WLI 1990, WP/WLI 1993). The four main types of movement are caused by
4. Results and Interpretation
the following factors:
Landslides are considered a natural process that occur in a variety of geologic settings,
1) Falls.
either These
as soil mass landslides
movement, involve
debristheflow,
collapse of materials
rockfall, from a cliff
or combinations or steep
of both. slope.
In this case,
Falls usually involve a mixture of free falls through the air, either bouncing
the main triggering factor has been the continuous agricultural irrigation of the crop fields or roll‐
ing. to
located A the
fall‐type
northlandslide results in Figure
(La Joya Antigua, the collection of rock
1). So, for or(around
years debris near
1968the base
until of a
today),
slope.
irrigation water has been infiltrating the ground through surface cracks and sink holes
2) Topples.
to saturate theTopple failures
subsurface. In involve the forward
fact, irrigation rotation
not only andlandslides
triggers movement in of
thisa mass of
area, but
rock, earth, or debris off a slope. This kind of slope failure generally
the occurrence of a landslide itself also increases the chances for new events to happen, asoccurs around
an axis
observed in(or
thepoint) at or near
reactivation the base
of 2016 of the
(Figure 2). block of rock.
3) Flows. Flows are landslides that involve the movement of material down a slope in
4.1. the
Basisform
for theofInterpretation
a fluid. Flows often leave behind a distinctive, upside‐down fun‐
nel‐shaped deposit
Landslides are classified wherebythe landslide
their material has
type of movement stopped
[41,42], and moving.
actually theThere are
scheme
different types of flows: mud, debris, and rock (rock avalanches).
terminology is suggested by the UNESCO Working Party on the ‘World Landslide Inven-
 4) Rotational and translational slides. Rotational slides occur on curved slip surfaces
where the upper surface of the displaced material may tilt backward toward the
scarp, whereas a translational (or planar) landslide is a downslope movement of
material that occurs along a distinctive planar surface of weakness, such as a fault,
joint, or bedding plane. Some of the largest and most damaging landslides on Earth
Geosciences 2023, 13, 342 9 of 19
are translational. These landslides occur at all scales and are not self‐stabilizing.
They can be very rapid when discontinuities are steep.
Figure 6 contains general schemes of the internal structure of a rotation‐
tory’ (WP WLI slide
al/transitional 1990,and
WP/WLI 1993).
the usual The four main
nomenclature types
of its main ofparts.
movementFrom are
these caused by theit
drawings,
following factors:that the main goal of this study is to determine the inner geometry–
can be inferred
geology.
(1) Falls.So, this means
These landslidesthat involve
the application of geophysical
the collapse of materials methods
from a is recommended
cliff or steep slope.to
develop
Fallsstructural and hydrogeological
usually involve a mixture of free inner
fallsmodels
through forthe
this
air,sliding
either mass.
bouncing or rolling.
ATaking thislandslide
fall-type classification
resultsinto account,
in the the of
collection Pierock
de or
Cuesta
debrislandslide
near the belongs
base of ato this
slope.
(2)
type.Topples.
Such cases Topple failures
are often veryinvolve the forward
susceptible rotationbecause,
to reactivation and movement
after theofslip
a mass of
has oc‐
rock,
curred, theearth, or debrisposition
equilibrium off a slope. This kind
is reached when of slope failuredecreases.
the torque generally So,
occurs
anyaround
changean in
axis (or point) at or near the base of the block of rock.
the water content or the removal of the lower part can lead to a new instability and thus,
(3) Flows. Flows
a reactivation areslip.
of the landslides that involve
Therefore, the movement
a diagnosis of the inner of material
geometrydown a slope
of these in the
phenom‐
ena isform
neededof a to
fluid.
make Flows
safetyoften leave behind a distinctive, upside-down funnel-shaped
estimations.
deposit
For thiswhere the landslide
purpose, the 2D‐ERT material has stopped
resulting modelsmoving. There are according
are interpreted different types of
to this
schemeflows:
andmud, debris, and rock information
the hydrogeological (rock avalanches).
[14,43,44], while the 2D‐SVP models focus
(4) Rotational
on detecting theand translational
compaction slides.
of these Rotationalmaterials,
sedimentary slides occur on curved
particularly slip surfaces
bedrock geom‐
where the upper surface of the displaced material may
etry. Figure 7 includes the correlation between the geologic materials (lithologic tilt backward toward the
column),
scarp, whereas a translational (or planar) landslide is a downslope
the resistivity ranges (Ωm), and the P‐wave velocities (m/s). Given that, low‐ resistivity movement of
valuesmaterial
correspondthat occurs along materials
to saturated a distinctive dueplanar
to the surface of weakness,(salinized)
high mineralization such as a fault,
of ir‐
joint,
rigation or bedding
water. plane. Some
Hydrological studiesof the largest
carried outand most
in the damaging
area indicatelandslides on Earth
water conductivity
are translational.
between 5 and 8 mS/cm These landslides
equivalent occur atΩm
to 1.25–2 all scales
[32]. On andtheareother
not self-stabilizing. They
hand, low velocities
can be very rapid when discontinuities are steep.
are related to soft materials, whereas an increasing velocity is proportional to their
Figure 6 contains general schemes of the internal structure of a rotational/transitional
compaction.
slide and the velocity
On the usual nomenclature
scale, 1500 of
m/sitshas
main parts.
been From these
marked as thedrawings, it canlevel
approximate be inferred
for the
that thetable.
water mainThis
goal value
of thiscorresponds
study is to determine the inner
to the velocity geometry–geology.
of P‐waves through theSo, this (under
water means
that the application of geophysical methods is recommended to develop structural
standard conditions) and in free or multilayer aquifers. As in the case of this landslide, and
hydrogeological inner models for this sliding mass.
this value indicates the level at which the unconsolidated sediments are saturated [32].

Figure 6.6. Left

Figure Left image
image shows
shows the
the inner
innergeometry
geometryofofthe
therotational/translational
rotational/translational landslide
landslide and
and its
its
nomenclature. Images on the right represent the different degrees of complexity that slippage
nomenclature. Images on the right represent the different degrees of complexity that slippage can can
have. Modified from Varnes 1996 [34].
have. Modified from Varnes 1996 [34].

Taking this classification into account, the Pie de Cuesta landslide belongs to this type.
Such cases are often very susceptible to reactivation because, after the slip has occurred,
the equilibrium position is reached when the torque decreases. So, any change in the
water content or the removal of the lower part can lead to a new instability and thus, a
reactivation of the slip. Therefore, a diagnosis of the inner geometry of these phenomena is
needed to make safety estimations.
Geosciences 2023, 13, 342 10 of 19

For this purpose, the 2D-ERT resulting models are interpreted according to this scheme
and the hydrogeological information [14,43,44], while the 2D-SVP models focus on de-
tecting the compaction of these sedimentary materials, particularly bedrock geometry.
Figure 7 includes the correlation between the geologic materials (lithologic column), the
resistivity ranges (Ωm), and the P-wave velocities (m/s). Given that, low- resistivity values
correspond to saturated materials due to the high mineralization (salinized) of irrigation
water. Hydrological studies carried out in the area indicate water conductivity between
Geosciences 2023, 13, x FOR PEER REVIEW 10 of 195
and 8 mS/cm equivalent to 1.25–2 Ωm [32]. On the other hand, low velocities are related to
soft materials, whereas an increasing velocity is proportional to their compaction.

(a) (b)
Figure 7. (a) Representative lithologic column of the study area. (b) Correlation between the geo‐
Figure 7. (a) Representative lithologic column of the study area. (b) Correlation between the
physical parameters and the lithology, hydrology, and compaction of the materials present in this
geophysical parameters and the lithology, hydrology, and compaction of the materials present in
landslide.
this landslide.
4.2. Interpretation of the 2D‐ERT Profiles
On the velocity scale, 1500 m/s has been marked as the approximate level for the
water The electrical
table. This valueprofiles were interpreted
corresponds according
to the velocity to thethrough
of P-waves two geological
the water targets
(under
where they were acquired. The ERT‐06 and ERT‐07 profiles are
standard conditions) and in free or multilayer aquifers. As in the case of this placed in the non‐sliding
landslide, this
area,
valuebehind
indicatesthethe
scarp.
levelAnd the other
at which five profiles (from
the unconsolidated ERT‐01
sediments aretosaturated
ERT‐05) are [32].located
within the landslide (Figure 3).
4.2. Interpretation
Concretely, the of the 2D-ERT
first Profilesare located in the crop fields. ERT‐06 is the closest
two profiles
to theThe
escarp, and ERT‐07 is parallel
electrical profiles were interpretedand 200according
m furthest to away.
the two Asgeological
the sediments
targetsinwhere
these
profiles
they were areacquired.
not sliding,TheitERT-06
was possible to establish
and ERT-07 profilesaarestratigraphic
placed in the correlation,
non-sliding andarea,
by
fixing
behindthe theboundaries
scarp. Andofthe three
othergeological formations,
five profiles the infiltration
(from ERT-01 to ERT-05) impact of irrigation
are located within
water on the materials
the landslide (Figure 3).was assessed. In both profiles (Figure 8a), the two water tables at
1492 Concretely,
m s.l. (Millothe Fm and
first twoLower Moquegua
profiles Fmincontact),
are located the cropand 1336ERT-06
fields. m s.l. is(Upper and
the closest
to the escarp,
Lower Moquegua and Fms
ERT-07 is parallel
contact) wereand 200 m furthest
marketed. With theseaway. As the sediments
divisions, it can be seenin these
that
profiles
there areare not sliding,
infiltration it was
areas in possible
the croptofields.
establish
Thea highest
stratigraphic correlation,
infiltration and byin
rate occurs fixing
the
the boundaries
northern of three geological
area, coinciding with theformations, the infiltration
side of the steepest impacttoward
scarp, while of irrigation waterthis
the south, on
the materials
infiltration was assessed.
decreases at the In both
ends profiles
(profile (Figure
ERT‐ 07) 8a),
at 950themtwo water
long, tables
where theatcrown
1492 mfin‐
s.l.
ishes.
Geosciences 2023, 13, 342 11 of 19

(Millo Fm and Lower Moquegua Fm contact), and 1336 m s.l. (Upper and Lower Moquegua
Fms contact) were marketed. With these divisions, it can be seen that there are infiltration
areas in the crop fields. The highest infiltration rate occurs in the northern area, coinciding
with the side of the steepest scarp, while toward the south, this infiltration decreases
Geosciences 2023, 13, x FOR PEER REVIEW 11 of 19 at the
ends (profile ERT- 07) at 950 m long, where the crown finishes.

(a)

(b)

Figure 8. Cont.
Geosciences 2023, 13, 342 12 of 19
Geosciences 2023, 13, x FOR PEER REVIEW 12 of 19

(c)
Figure 8. (a) Interpretation of the electric profiles placed outside the landslide (ERT‐06 and ERT‐07,
Figureon (a) Interpretation
8. previous page). (b,c) of the electricofprofiles
Interpretation placed
the electric outside
profiles placedthe landslide
within (ERT-06
the landslide and ERT-07, on
(from
previous page). (b,c) Interpretation of the electric profiles placed within the landslide (from ERT-01
ERT‐01 to ERT‐05).
to ERT-05).
In general, the conglomerate is a permeable lithology that has a high resistivity be‐
tween 200 and 800 Ωm (depending on the amount of interstitial water and its composi‐
Intion).
general, the conglomerate is a permeable lithology that has a high resistivity be-
In this case, the low resistivities detected in the Upper Moquegua Fm indicate that
tweenthese
200 andsedimentsΩm
800 have(depending
saline water onandthe amount
resistivity of interstitial
values lower than water
10–20 Ωm and(blue),
its composition).
In thiswhich
case,suggests
the lowtheresistivities detectedsectors.
existence of saturated in theThe Upper
resultMoquegua
is an interfaceFm indicate
of high hu‐ that these
midityhave
sediments that saline
can reactivate
water the andlandslide
resistivity if any instability
values lower occurs.
thanThis
10–20is inΩmaccordance
(blue), which sug-
with the 2015–2019 years of reactivation, where the crown was mostly displaced in the
gests the existence of saturated sectors. The result is an interface of high humidity that can
most conductive zones (Figure 2). Another interesting aspect is the plume shape of the
reactivate the landslide
two main if anyininstability
infiltration zones; both cases, they occurs. This
appear is in
to be accordance
flowing toward the with
Lower the 2015–2019
years of reactivation,
Moquegua where
Fm contact. the
If we crown
look was
at profile mostly
ERT‐06, thedisplaced in the
section between 260most
and 460 conductive
m is zones
(Figurethe2).most affected
Another for water plume,
interesting aspectand is theconsidering the proximity
plume shape to themain
of the two crown, this
infiltration zones;
in both segment
cases,could
theybeappear
the trigger
to for
be another
flowing future reactivation.
toward the Lower Moquegua Fm contact. If we
The profiles within the landslide show different aspects of the internal structure of
look attheprofile ERT-06, the section between 260 and 460 m is the most affected for water
landslide depending on where it was made (Figure 8b). In this way, the ERT‐01 and
plume,ERT‐02
and considering the in
profiles are placed proximity
the upper part,to the crown,
below thisscarp
the main segment
(Figure could be are
2). They the trigger for
another future
nearly reactivation.
parallel and run longitudinally to the slipped materials. In them, we can see a
sliding
The mass within
profiles characterized by low resistivities
the landslide show different (lower aspects
than 10–20 Ωm,internal
of the blue values)structure of the
bounded by the slip plane (rupture surface), which defines the top of the dry basement
landslide depending on where it was made (Figure 8b). In this way, the ERT-01 and ERT-02
with high resistivities (higher than 200 Ωm, yellow‐red values). At the end of these two
profiles are placed
profiles, in the
we can see upper part,
the elevation below to
corresponding thethemain
toe of scarp (Figure
the rupture 2). where
surface, They are nearly
parallelthe debris (remains dray) accumulates and a system of cracks is detected on both sidesaofsliding mass
and run longitudinally to the slipped materials. In them, we can see
this elevation.
characterized by low Thisresistivities
relief is distinctive
(lowerfor thanthe10–20 Ωm, blue values)
rotational/transitional landslide,
boundedafter by the slip
which the fluidization zone begins. The ERT‐01 profile
plane (rupture surface), which defines the top of the dry basement with high has lower resistivities for the toe resistivities
rupture surface (between 60 and 160 Ωm, yellow‐green values), suggesting that the ma‐
(higher than 200 Ωm, yellow-red values). At the end of these two profiles, we can see
terials may have a higher mineralized water content. It is precisely this profile that is lo‐
the elevation
cated in thecorresponding
most advanced lobe to the toe
of the of the rupture surface, where the debris (remains
slide.
dray) accumulates
Longitudinal and a system
profiles ERT‐04of cracks
and ERT‐05 is(Figure
detected on placed
8c) are both sides oftoe
after the thisof elevation.
the This
rupture
relief is surface, and
distinctive they rotational/transitional
for the show the sliding mass in thelandslide, fluidization zone,
after where
which thethede‐fluidization
tected resistivities are slightly higher (around 30 Ωm, light blue values). This aspect can be
zone begins. The ERT-01 profile has lower resistivities for the toe rupture surface (between
interpreted in two ways: i) as a loss of water content of the materials because they have
60 andcome Ωm,
160into yellow-green
contact values),
with river deposits; ii) as suggesting
a mixture withthat thewater
the fresh materials may have a higher
of the valley.
mineralized water content. It is precisely this profile that is located in the most advanced
lobe of the slide.
Longitudinal profiles ERT-04 and ERT-05 (Figure 8c) are placed after the toe of the
rupture surface, and they show the sliding mass in the fluidization zone, where the detected
resistivities are slightly higher (around 30 Ωm, light blue values). This aspect can be
interpreted in two ways: (i) as a loss of water content of the materials because they have
come into contact with river deposits; (ii) as a mixture with the fresh water of the valley.
Finally, profile ERT-03 (Figure 8b) was acquired transversally behind the toe rupture
surface in the fluidization zone. In this case, it is observed that the sliding mass is more
resistive in the central lobe, coinciding with the two previous profiles, while at the NE,
Geosciences 2023, 13, 342 13 of 19

Geosciences 2023, 13, x FOR PEER REVIEW 13 of 19

under the second advancing lobe, it is more conductive. This could suggest a greater
plasticity of materials, which would give this lobe a higher risk of advancing.
Finally, profile ERT‐03 (Figure 8b) was acquired transversally behind the toe rupture
surface in the fluidization zone. In this case, it is observed that the sliding mass is more
4.3. Interpretation
resistive inofthe
thecentral
2D-SVP Seismic Profiles
lobe, coinciding with the two previous profiles, while at the NE,
under the seismic
The SVP-01 second advancing
profilelobe, it is more
(Figure 3) conductive.
is planned Thisalmost
could suggest a greater
coincidentally with the
plasticity of materials, which would give this lobe a higher risk of advancing.
transversal electrical profile ERT-03 and SVP-02 with the longitudinal ERT-01 profile. In
Figure 9, both electricalof and
4.3. Interpretation seismic
the 2D‐SVP equivalent
Seismic Profiles models have been jointly interpreted. In
the SVP-01 model, we seismic
The SVP‐01 can see three
profile velocity
(Figure layersalmost
3) is planned (Figure 9a). Thewith
coincidentally firsttheone is a thick
transversal electrical profile ERT‐03 and SVP‐02 with the longitudinal ERT‐01 profile. In
layer (200 m) made up of loose materials with a low velocity (Vp < 1400 m/s) that can be
Figure 9, both electrical and seismic equivalent models have been jointly interpreted. In
associatedthe
with the model,
SVP‐01 sliding wemass
can seedetected in the
three velocity ERT-03
layers (Figureprofile. Below
9a). The first one this, an intermediate
is a thick
layer has been detected
layer (200 m) made related to moderately
up of loose materials with astiff materials
low velocity (Vp <comprising
1400 m/s) that between
can be 1500 m/s
associated with the sliding mass detected in the ERT‐03 profile. Below this, an interme‐
and 3000 m/s. These values have been set in ERT-03 like respective iso-velocity lines (white
diate layer has been detected related to moderately stiff materials comprising between
and red lines
1500inm/sFigure
and 3000 9a),
m/s.and
These we canhave
values infer that
been set the 1500like
in ERT‐03 m/s value iso‐velocity
respective marks the water level
(approx.), while 3000 m/s defines the top of the strong materials linked
lines (white and red lines in Figure 9a), and we can infer that the 1500 m/s value with the Lower
marks
the water level (approx.), while 3000 m/s defines the top of
Moquegua Fm. On the NE side of both models (SRT-01 and ERT-03), this middle layer canthe strong materials linked
with the Lower Moquegua Fm. On the NE side of both models (SRT‐01 and ERT‐03), this
be correlated
middlewith the
layer sliding
can mass,with
be correlated while at themass,
the sliding SW while
end, at it the
seems to coincide
SW end, it seems towith the river
deposits (materials
coincide with ofthe
the valley).
river depositsNote thatofthe
(materials the contact between
valley). Note that the these
contact two geological envi-
between
ronments isthese two geological
described by aenvironments
local reliefis (200
described
m) by a local
that reliefact
could (200asm)athat could act as
“barrier”. Ina Figure 10b, it
“barrier”. In Figure 10b, it can be seen how this relief coincides with the end of the toe
can be seen how this relief coincides with the end of the toe subsurface.
subsurface.

Geosciences 2023, 13, x FOR PEER REVIEW 14 of 19

(a)

(b)
Figure 9. Interpretation of the seismic profiles placed within the landslide. (a) The SVP‐01 is almost
Figure 9. Interpretation
coincident withof the
theERT‐03
seismic profiles
profile placed
and their withinhasthe
interpretation beenlandslide. (a) Theseismic
done by correlating SVP-01 is almost
coincident withandthe ERT-03
electrical profile
responses. (b)and theirprocedure
The same interpretation hasfor
has been used been done by correlating seismic and
the SVP02.

electrical responses. (b) The same procedure has been used for the SVP02.
The beginning of the SVP‐02 seismic profile (Figure 9b) coincides with the end of the
main scarp over the heat sector (Figures 3 and 6). In this place, the upper layer has a
well‐defined velocity of 1500 m/s, indicating that there is a significant inflow of water.
The top basement rock is also clear (4000 m/s), and there is hardly any middle layer.
Around a distance of 150 m, we can see an inner fault that would correspond structurally
to the displacement of the minor escarp, characterized by a displacement of about 35 m
(jump). If we correlated this seismic profile with the ERT‐01 electric profile, the materials
 images above correspond to the DTM with a resolution of 2m/pixel (Figure 11a) and the
deduced 3D rupture surface model without the electrical profiles (Figure 11b), giving
1,777,000 m2 for a no‐planar area and a 1,247,748 m2 for a planar area. On this subsurface,
the preferred runoff directions set out in Figure 10 have been placed. The model pre‐
Geosciences 2023, 13, 342
sented in Figure 11c is a composite that demonstrates the sliding subsurface related 14
toofthe
19
DTM, where each node of the mesh preserves its UTM coordinates; with the vertical (z)
expressed in m above sea level (m.a.s.l., altitudes).

Figure 10. (a) A 3D layout of 2D‐ERT electrical models below the digital elevation model. (b) Ex‐
Figure 10. (a) A 3D layout of 2D-ERT electrical models below the digital elevation model.
trapolated top of basement roc (Lower Moquegua Fm.) from the 2D‐ERT interpretation. The arrows
(b) Extrapolated
indicate preferredtop of basement
flow directions.roc (Lower Moquegua Fm.) from the 2D-ERT interpretation. The
arrows indicate preferred flow directions.
If we compare this sliding subsurface with the graph of the inner geometry for the
The beginning of the SVP-02 seismic profile (Figure 9b) coincides with the end of
rotational/translational landslide in Figure 6, we can establish the main parts for the Pie
the main scarp over the heat sector (Figures 3 and 6). In this place, the upper layer has a
de Cuesta slide: (i) the upper heat zone comprises altitudes between 1500 and 1300 m
well-defined velocity of 1500 m/s, indicating that there is a significant inflow of water. The
(first thick contour line), (ii) the minor scarp is placed at depths of 1300–1230 m (second
top basement rock is also clear (4000 m/s), and there is hardly any middle layer. Around
thick contour line), and (iii) the toe surface rupture is the “relief” bordering on the north
a distance of 150 m, we can see an inner fault that would correspond structurally to the
of the main lobe. It is noteworthy that the toe ends just behind the main lobe of the sliding
displacement of the minor escarp, characterized by a displacement of about 35 m (jump).
and is not detected in the borderline profile ERT‐04. This may be interpreted as a line of
If we correlated this seismic profile with the ERT-01 electric profile, the materials at the
interception, in this case, capping between the bottom of the rupture surface and the
bottom of the rupture surface between 175 and 450 m long are moderately stiff and clear
original
(3000 m/s),ground surface.
coinciding Above
with it, the sliding
medium-values mass flows with the fluidized materials.
resistivities.
The seismic profile SRT‐02 marks velocities of 500–1500 m/s for the involved materials.
4.4. 3D Subsurface Models
In order to obtain global knowledge of the Pie de Cuesta landslide, all geophysical
profiles have been georeferenced and placed under the Digital Terrain Model (DTM)
(Figure 10a). In this figure, we used only the ERT profiles to present a clear image. Once
the profiles have been spatially distributed (GNSS georeferenced), the rupture surface has
been deduced by digitizing the top of the resistive basement roc (Lower Moquegua Fm.) in
each profile and extrapolating the point values for the landslide sector that has a reasonable
geophysical coverage (Figure 10b). As described in the previous section (Figure 6), this
digital and georeferenced rupture surface represents the subsurface where the sliding mass
circulates and approximately coincides with the top of the strong roc.
Geosciences 2023, 13, 342 15 of 19

Taking into account the fact that the upper plain corps (La Joya Antigua) is irrigated
with salinized water that produces conductive effects, and considering the fact that when
this water infiltrates, it follows along the maximum slopes of the rupture surface (at similar
lithology), a first attempt has been made to establish the main flow entrances and their
directions. Figure 10b shows the flow directions of water infiltrations; they are detected
mainly in the sectors close to the lateral scarps and, apparently, with a greater water
contribution on the north side, coinciding with the prominent toe of the subsurface.
Likewise, the 3D models in Figure 11 are the final results of the entire study. The
two images above correspond to the DTM with a resolution of 2m/pixel (Figure 11a) and
Geosciences 2023, 13, x FOR PEER REVIEW
the deduced 3D rupture surface model without the electrical profiles (Figure 11b), 16 of 19
giving
2 2
1,777,000 m for a no-planar area and a 1,247,748 m for a planar area. On this subsurface,
the preferred runoff directions set out in Figure 10 have been placed. The model presented
in Figure 11c11c
Figure is a demonstrates
composite thatthe demonstrates the sliding
difference between subsurface
the MDT andrelated to thesliding
the entire DTM,
where each node
subsurface and of the mesh the
represents preserves its UTMsliding
accumulated coordinates;
mass with thethe
along vertical
major(z)part
expressed
of the
in m abovegiving
landslide, sea level (m.a.s.l.,
a total volumealtitudes).
of 90,570, 516 m3 (about 90 Hm3).

Figure11.
Figure 11. (a)
(a)Digital
DigitalTerrain
TerrainElevation
Elevationmodel
model(DTM)
(DTM)of
ofthe
thePie
Piede
deCuesta
Cuestalandslide.
landslide.(b)
(b)Deduced
Deduced
rupture surface from the geophysical methods. (c) Rupture surface under DTM. (d) Accumulated
rupture surface from the geophysical methods. (c) Rupture surface under DTM. (d) Accumulated
sliding mass.
sliding mass.

5. Discursion and Conclusions

The objective of this work has been to present the methodology followed in the
study of a large landslide, as in the case of the active Pie de Cuesta slide. In all cases, a
previous geological study is decisive in recognizing the lithologies present and under‐
Geosciences 2023, 13, 342 16 of 19

If we compare this sliding subsurface with the graph of the inner geometry for the
rotational/translational landslide in Figure 6, we can establish the main parts for the Pie
de Cuesta slide: (i) the upper heat zone comprises altitudes between 1500 and 1300 m
(first thick contour line), (ii) the minor scarp is placed at depths of 1300–1230 m (second
thick contour line), and (iii) the toe surface rupture is the “relief” bordering on the north of
the main lobe. It is noteworthy that the toe ends just behind the main lobe of the sliding
and is not detected in the borderline profile ERT-04. This may be interpreted as a line
of interception, in this case, capping between the bottom of the rupture surface and the
original ground surface. Above it, the sliding mass flows with the fluidized materials. The
seismic profile SRT-02 marks velocities of 500–1500 m/s for the involved materials.
Figure 11c demonstrates the difference between the MDT and the entire sliding sub-
surface and represents the accumulated sliding mass along the major part of the landslide,
giving a total volume of 90,570, 516 m3 (about 90 Hm3 ).

5. Discursion and Conclusions

The objective of this work has been to present the methodology followed in the study
of a large landslide, as in the case of the active Pie de Cuesta slide. In all cases, a previous
geological study is decisive in recognizing the lithologies present and understanding
their behavior when they are saturated. In this case, the geological survey consisted of
defining the geomorphology and the stratigraphic series and determining that it is mainly
a rotational/translational type of landside. At present, this landslide is “almost inactive”,
but there is a high probability of reactivation because any small change in the water content
or removal of the lower part can lead to a new great instability.
In this situation, the geophysical profiles that have been carried out provide valu-
able information on the internal geometry of the landslide. According to the results, the
following conclusions can be drawn:
(i) With the electrical profiles, it has been possible to discriminate the sliding (conduc-
tive) mass from the resistive bedrock. Between these two layers, we identified the transition
zone with medium resistivity values, which we related to a level of debris and the erosional
top of the basement.
Another important contribution of electrical exploration has been that it is possible to
establish, albeit only approximately, the preferred directions of water flow (by infiltrations
or runoffs); according to the principle that for the same material, a decrease in resistivity
can be associated with an increase in water content, especially when the impregnation
water is highly mineralized, as in this case.
(ii) Although only two seismic profiles were made for the test, they have shown that
when there is a good contrast between the densities of the materials involved, the obtained
models provide good information on their compaction and degree of geotechnical safety.
When comparing the slip masses detected in seismic and electrical surveys (coinci-
dent), we can see that they coincide well when the top basement is a compact or aquitard
formation. However, there are cases where the top basement is eroded, fractured, or
partially dissolved, and the water infiltrates, reducing resistivities, while velocities are
little affected (in the second order). Then, when the models disagree and only id electrical
data are available, the indeterminacy arises. In the Pie de Cuesta landslide, there is the
possibility of this happening in some sectors because the basement roc (Lower Moquegua
Fm.) would have layers of gypsum on its top and below that would be very compact and
dry clays. So, the levels of gypsum may be partially impregnated and saturated, and the
resistivity decreases at a greater rate than the velocity, whereas in compacted dry clays,
the velocities and resistivities remain high. So, the integration of all the data enabled us to
better identify the geometry of the landslide.
With the 3D rupture surface model, it is possible to know the inner “landslide geome-
try”, the main flow directions, and calculate the volume and distribution of the sliding mass.
These results provide valuable information to gusset a geotechnical diagnostic. Having
established this critical slip surface for a large area of the landslide gives an advantage
Geosciences 2023, 13, 342 17 of 19

of the geophysical survey over borehole data, both for its high cost and for its very local
information. However, in a general study, both techniques are crucial.
Finally, we consider this case study an interesting example of geophysical exploration
on large landsides. The difficult topography caused by the multiple slumps and the
thickness of the sliding mass present a significant handicap to placing the geophysical
profiles. This is a crucial aspect because it greatly depends on the accessibility of the
area and the availability of the physical space required. In our case, we need to extend
profiles up to 1100 m long in order to obtain data at greater depths since this landslide is
approximately 200 m tall.
Our last comment concerns the validity of the 3D geophysical model obtained. The
most common way to verify the geophysical results is to check them using well logs.
Although the boreholes only provide good information at one subsurface point, they are a
useful tool for calibrating the geophysics profiles at their intersection point. This ensures
the reliability of the rest of the profile. Unfortunately, no sounding has yet been carried
out on this site to assess the degree of accuracy, and our geophysical survey has only been
checked in the areas where a geological inspection of the “lateral-depths” could be carried
out. mainly in lateral slopes and the final toe part.

Author Contributions: Conceptualization, J.S. and K.L.; methodology, T.T. and K.L.; software, Y.H.
and K.L.; validation, T.T., K.L. and J.S.; resources, J.H., R.I. and Y.A.; writing—review and editing,
Y.H., R.I., J.S., J.H., Y.A. and T.T.; supervision, J.S., K.L. and J.H.; All authors have read and agreed to
the published version of the manuscript.
Funding: This research was funded by the Perú National Research Program funding the National
University of San Agustín of Arequipa to carry out the Research Project grant number IBAIB-03-2018-
UNSA. The APC was funded by the National University of San Agustín of Arequipa.
Data Availability Statement: Data are contained within the article.
Acknowledgments: Carlos Araque, for his important contribution in the development of 3D mod-
els for this research. To Fredy Perez, for his important support in field work and processing of
seismic lines.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Anis, Z.; Wissem, G.; Riheb, H.; Biswajeet, P.; Essghaier, G.M. Effects of clay properties in the landslides genesis in flysch massif:
Case study of Aïn Draham, North Western Tunisia. J. Afr. Earth Sci. 2019, 151, 146–152. [CrossRef]
2. Clow, D.W.; Schrott, L.; Webb, R.; Campbell, D.H.; Torizzo, A.; Dorblaser, M. Ground water occurrence and contributions to
streamflow in an alpine catchment, Colorado Front range. Groundwater 2003, 41, 937–950. [CrossRef]
3. Garcia-Chevesich, P.; Wei, X.; Ticona, J.; Martínez, G.; Zea, J.; García, V.; Alejo, F.; Zhang, Y.; Flamme, H.; Graber, A.; et al. The
Impact of Agricultural Irrigation on Landslide Triggering: A Review from Chinese, English, and Spanish Literature. Water 2021,
13, 10. [CrossRef]
4. Perrone, A.; Lapena, V.; Piscitelli, S. Electrical resistivity tomography technique for landslide investigation: A review. Earth-Sci.
Rev. 2014, 135, 65–82. [CrossRef]
5. Bichler, A.; Bobrowsky, P.; Best, M.; Douma, M.; Hunter, J.; Calvert, T.; Burns, R. Three-dimensional mapping of a landslide using
a multi-geophysical approach: The Quesnel Forks landslide. Landslides 2004, 1, 29–40. [CrossRef]
6. Jongmans, D.; Hemroulle, P.; Demanet, D.; Renardy, F.; Vanbrabant, Y. Application of 2D electrical and seismic tomography
techniques for investigating landslides. Eur. J. Environ. Eng. Geophys. 2000, 5, 75–89.
7. McCann, D.M.; Forster, A. Reconnaissance geophysical methods in landslide investigations. Eng. Geol. 1990, 29, 59–78. [CrossRef]
8. Bruno, F.; Marillier, F. Test of high-resolution seismic reflection and other geophysical techniques on the Boup landslide in the
Swiss Alps. Surv. Geophys. 2000, 21, 333–348. [CrossRef]
9. Mauritsch, H.J.; Seiberl, W.; Arndt, R.; Romer, A.; Schneiderbauer, K.; Sendlhofer, G.P. Geophysical investigations of large
landslides in the Carnic Region of southern Austria. Eng. Geol. 2000, 56, 373–388. [CrossRef]
10. Lapenna, V.; Lorenzo, P.; Perrone, A.; Piscitelli, S.; Sdao, F.; Rizzo, E. High-resolution geoelectrical tomographies in the study of
the Giarrossa landslide (Potenza, Basilicata). Bull. Eng. Geol. Environ. 2003, 62, 259–268. [CrossRef]
11. Hood, J.L.; Roy, J.W.; Hayashi, M. Importance of groundwater in the water balance of an alpine headwater lake, Geophys. Res.
Lett. 2006, 33, L13405. [CrossRef]
Geosciences 2023, 13, 342 18 of 19

12. Garrido, J.; Delgado, J. A recent, retrogressive, complex earthflow–earth slide at Cenes de la Vega, southern Spain. Landslides
2013, 10, 83–89. [CrossRef]
13. Godio, A.; Bottino, G. Electrical and electromagnetic investigation for landslide characterization. Phys. Chem. Earth Part C Sol.
Terr. Planet. Sci. 2001, 26, 705–710. [CrossRef]
14. Shevnin, V.; Delgado-Rodríguez, O.; Mousatov, A.; Ryjov, A. Estimation of hydraulic conductivity on clay content in soil
determined from resistivity data. Geofis. Int. 2006, 45, 195–207. [CrossRef]
15. Khalil, M.A.; Monterio-Santos, F.A. Influence of Degree of Saturation in the Electric Resistivity–Hydraulic Conductivity Relation-
ship. Surv. Geophys. 2009, 30, 601–615. [CrossRef]
16. Sestras, P.; Bilas, co, S, .; Ros, ca, S.; Veres, I.; Ilies, N.; Hysa, A.; Spalević, V.; Cîmpeanu, S.M. Multi-Instrumental Approach to
Slope Failure Monitoring in a Landslide Susceptible Newly Built-Up Area: Topo-Geodetic Survey, UAV 3D Modelling and
Ground-Penetrating Radar. Remote Sens. 2022, 14, 5822. [CrossRef]
17. U.S. Geological Survey. Landslide Hazards Program. Available online: https://www.usgs.gov/programs/landslide-hazards
(accessed on 29 October 2023).
18. Sass, O. Bedrock detection and talus thickness assessment in the European Alps using geophysical methods. J. Appl. Geophys.
2007, 62, 254–269. [CrossRef]
19. Abolmasov, B.; Ristic, A.; Govedaric, M. Applying GPR and 2D ERT for shallow landslides characterization: A case study. In
Landslide Science and Practice; Margottini, C., Canuti, P., Sassa, K., Eds.; Springer: Berlin/Heidelberg, Germany, 2013; Volume 2.
[CrossRef]
20. Lghoul, M.; Teixidó, T.; Peña, J.A.; Hakkou, R.; Kchikach, A.; Guerin, R.; Jaffal, M.; Zouhri, L. Electrical and Seismic Tomography
Used to Image the Structure of a Tailings Pond at the Abandoned Kettara Mine, Morocco. Mine Water Environ. 2012, 31, 53–61.
[CrossRef]
21. Cahyaningsih, C.; Riau, U.I.; Asteriani, F.; Crensonni, P.F.; Choanji, T. Safety factor characterization of landslide in Riau-West of
Sumatra highway. Int. J. Geomate 2019, 17, 323–330. [CrossRef]
22. Himi, M.; Anton, M.; Sendrós, A.; Abancó, C.; Ercoli, M.; Lovera, R.; Deidda, G.P.; Urruela, A.; Rivero, L.; Casas, A. Application of
Resistivity and Seismic Refraction Tomography for Landslide Stability Assessment in Vallcebre, Spanish Pyrenees. Remote Sens.
2022, 14, 6333. [CrossRef]
23. Huamán, G.E.A.; Pinto, W.P.; Chulla, H.L.O.; Cárdenas, J.H. Geodynamics, Geodetic Monitoring and Geophysical Prospection of the Pie
de Cuesta Landslide-Vítor, Arequipa; Mining and Metallurgical Geological Institute-INGEMMET: Lima, Peru, 2018.
24. Mining and Metallurgical Geological Institute. Monitoring of the Pie de Cuesta Landslide during the 2021 Period, Vitor and La Joya
District, Province of Arequipa, Department of Arequipa; Technical Report A7242; INGEMMET: Lima, Peru, 2022; p. 26.
25. Araujo, G.E.; Miranda, R. Geological and Geodynamic evaluation of landslides on the left flank of the Vitor Valley, Pie de Cuesta, Telaya,
Gonzales and Socabón Sectors. Vitor and La Joya Districts, Arequipa Region, Arequipa Province; Mining and Metallurgical Geological
Institute-INGEMMET: Lima, Peru, 2016.
26. Lacroix, P.; Dehecq, A.; Taipe, E. Irrigation-triggered landslides in a Peruvian desert caused by modern intensive farming. Nat.
Geosci. 2020, 13, 56–60. [CrossRef]
27. Huanca, J. Application of the GNSS-RTK Technique and Photogrammetry with Drones for the Characterization and Monitoring
of the Pie de Cuesta Active Landslide in the Vitor Valley, Arequipa. Licentiate Thesis, National University of San Agustín of
Arequipa, Arequipa, Peru, 2020.
28. Huerta, V. Rehabilitation of the Mocoro Canal—San Luis in the Pie de la Cuesta Section—La Cano Irrigation—Geological and
Geotechnical Study. Licentiate Thesis, National University of San Agustín of Arequipa, Arequipa, Peru, 1977.
29. Sass, O. Determination of the internal structure of alpine talus deposits using different geophysical methods (Lechtaler Alps,
Austria). Geomorphology 2006, 80, 45–58. [CrossRef]
30. Pazzi, V.; Morelli, S.; Fanti, R. A Review of the Advantages and Limitations of Geophysical Investigations in Landslide Studies.
Int. J. Geophys. 2019, 2019, 2983087. [CrossRef]
31. Emily, E. Brodsky, Evgenii Gordeev, Hiroo Kanamori. Landslide basal friction as measured by seismic waves. Geophys. Res. Lett.
2003, 30, 2236. [CrossRef]
32. Powers, M.H.; Burton, B.L. Questa Baseline and Pre-Mining Ground-Water Quality Investigation. Seismic Refraction Tomography for
Volume Analysis of Saturated Alluvium in the Straight Creek Drainage and Its Confluence with Red River, Taos County, New Mexico:
U.S.; Geological Survey Scientific Investigation Report 2006–5166; U.S. Geological Survey: Reston, VA, USA, 2007; p. 19,
ISBN 978-94-6282-119-4. [CrossRef]
33. Griffiths, D.H.; Barker, R.D. Two-dimensional resistivity imaging and modelling in areas of complex geology. J. Appl. Geophys.
1993, 29, 211–226. [CrossRef]
34. Edwards, L.S. A modified pseudosection for resistivity and IP. Geophysics 1977, 42, 1020–1036. [CrossRef]
35. Aki, K.; Richards, P.G. Quantitative Seismology: Theory and Methods; W. H. Freeman & Company: San Francisco, CA, USA, 1980;
Volume I, p. 557, ISBN 0716710587.
36. Samouëlian, A.; Cousin, I.; Tabbagh, A.; Bruand, A.; Richard, G. Electrical resistivity survey in soil science: A review. Soil Tillage
Res. 2005, 83, 173–193. [CrossRef]
37. Loke, M.H.; Barker, R.D. Rapid least-squares inversion of apparent resistivity pseudosections by a quasi-Newton method. Geophys.
Prospect. 1996, 44, 131–152. [CrossRef]
Geosciences 2023, 13, 342 19 of 19

38. Loke, M.H. 2-D and 3-D inversion and Modeling of Surface- and Borehole Resistivity Data; United States Geological Survey: Hartford,
CT, USA, 2001.
39. Tarantola, A.; Valette, B. Generalized Nonlinear inverse problems solved using the Least Squares Criterion. Rev. Geophys. 1982, 20,
219–232. [CrossRef]
40. Gebrande, H.; Miller, H. The DeltatV 1D Method for Seismic Refraction Inversion: Theory; II 226–260. Siegfried R. Rohdewald, 2011;
Intelligent Resources Inc.: Manchester, UK, 1985.
41. Varnes, D.M. Landslide Types and Processes, Transportation Research Board, U.S.; National Academy of Sciences, Special Report;
Transportation Research Board: Washington, DC, USA, 1996; Volume 247, pp. 36–57.
42. Hungr, O.; Leroueil, S.; Picarelli, L. The Varnes classification of landslide types, an update. Landslides 2014, 11, 167–194. [CrossRef]
43. McInnis, D.; Silliman, S.; Boukari, M.; Yalo, N.; Orou-Pete, S.; Fertenbaugh, C.; Sarre, K.; Fayomi, H. Combined application of
electrical resistivity and shallow groundwater sampling to assess salinity in a shallow coastal aquifer in Benin. West Afr. J. Hydrol.
2013, 505, 335–345. [CrossRef]
44. Lesmes, D.; Friedman, S. Relationships between the Electrical and Hydrogeological Properties of Rocks and Soils. In Hydrogeo-
physics; Springer: Berlin/Heidelberg, Germany, 2005. [CrossRef]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.