Uni ve r s i t at de Bar c e l ona - Fac ul t ad de Ge ol ogí a - De par t ame nt o de Ge odi námi c a y Ge of í s i c a

Pr ogr ama de doc t or ado e n Ci e nc i as de l a Ti e r r a - Bi e ni o 2004 - 2006
Gr up de r e ce r c a e n r i s c os nat ur al s RI SKNAT



Improvementsinourunderstandingofrockfallphenomenon
byTerrestrialLaserScanning
-
Emphasisonchangedetectionand
itsapplicationtospatialprediction

____________________________________________



Antonio Abellán Fernández


Supervisor:
Dr. Joan Manuel Vilaplana Fernández


Barcelona, July 2009







































The financial support of the Spanish Ministry of Science and Education (pre-doctoral grant 2004-1852) is gratefully
acknowledged. This work was funded by the Geomodels Institute, the Natural Park of the Garrotxa Volcanic Field and
following projects: MEC CGL2006-06596 (DALMASA) and TopoIberia CSD2006-0004/Consolider-Ingenio2010.



A mi familia y amigos
Con mención especial para B.



































































































PREFACE

This PhD thesis describes research carried out by Antonio Abellán Fernández as part of the Earth
Sciences Doctoral Program in the Geodynamics and Geophysics Department (Faculty of Geology,
University of Barcelona) during the period 2004 - 2009. The research was supervised by Dr. Joan
Manuel Vilaplana Fernández, head of the RISKNAT group in the Geodynamics and Geophysics
Department, University of Barcelona. Part of this research was carried out in collaboration with Dr.
Nicholas J . Rosser, professor at the Institute of Hazard and Risk Research, Geography Department,
University of Durham, UK and Dr. Michel J aboyedoff, head of the Institute of Geomatics and Risk
Analysis (IGAR), Faculty of Geosciences and Environment (GSE), University of Lausanne.

The core of this dissertation is a compendium of four research papers, of which two have been
published, one has been accepted and one has been submitted for publication.

i. Publication A (chapter 2): Abellán, A., Vilaplana, J.M., Martínez, J., 2006. Application
of a long-range Terrestrial Laser Scanner to a detailed rockfall study at Vall de Núria
(Eastern Pyrenees, Spain). Engineering Geology, 88(3-4), 136-148.

ii. Publication B (chapter 4): Abellán, A., Vilaplana, J.M., Calvet, J., García-Sellés, D.
Asensio, E., (submitted to NHESS). Rockfall monitoring by Terrestrial Laser Scanning.
Case study of the basaltic rock face at Castellfollit de la Roca (Catalonia, Spain).

iii. Publication C (chapter 5): Abellán, A. Jaboyedoff, M., Oppikoffer, T., Vilaplana, J.M.,
2009. Detection of millimetric deformation using a terrestrial Laser scanner: Experiment
and application to a rockfall event. Natural Hazards and Earth System Science, 9, 365-
372.

iv. Publication D (chapter 6): Abellán, A., Vilaplana, J.M., Calvet, J., Blanchard, J.
(accepted in Geomorphology): Detection and spatial prediction of rockfalls by means of
terrestrial laser scanning monitoring.































ABSTRACT

Minor-scale rockfalls (up to 100 m
3
) are the most frequent type of landslides on steep slopes,
such as rock faces, coastal cliffs and the sides of transportation corridors (Guzzetti et al., 2004;
Copons and Vilaplana, 2008; Lim et al., 2009). Since their impact energy can be very high
(Agliardi and Crosta, 2003; Dorren and Seijmonsbergen, 2004), rockfall phenomena present a
common risk to infrastructures, buildings and/or populations in steep mountainous terrain. The
overall aim of this PhD is to improve our understanding of rockfall phenomena by means of
terrestrial Light Detection and Ranging (LIDAR) technology. Specifically, we used a
Terrestrial Laser Scanner (TLS), ILRIS-3D model (Optech). This new technology acquires
dense 3D information of the terrain, with high accuracy (σ = 0.72cm at 50m), high resolution
(mm-cm order) and very high data acquisition speeds (up to 2,500 points/second). The TLS
instrument is currently being used in different applications related to rockfalls, such as the
generation of a high-resolution, high-accuracy Digital Elevation Models (DEM) of steep slopes,
the characterization of 3D discontinuities that play a key role in rockfall detachment, the
monitoring of rock slopes, the study of rockfall magnitude-frequency relationships, etc. (e.g.
Slob and Hack, 2004; Bauer et al., 2005; Rosser et al., 2005; Lim et al., 2006; Jaboyedoff et al.,
2007; Oppikoffer et al, 2008a; Sturzenegger and Stead, 2009). In relation to these applications,
one of the most interesting challenges is the possibility to link spatial and temporal prediction of
rockfalls (Rosser et al., 2007; Abellán et al., 2009; Abellán et al., accepted).

The selected study areas are characterized by highly variable geomorphological scene,
lithologies, geomechanical behaviour, hazard and risk levels, etc. The areas were: (i) a sector of
the steep gneissic slopes at Vall de Núria (Publication A); (ii) the basaltic cliff at Castellfollit de
la Roca (Publication B); and (iii) the main scarp of an old landslide at Puigcercós, mainly
formed by marls and sandstones (Publication D). In addition, an experimental test in artificial
scenery (Publication C) allowed the instrumental and methodological issues to be better
understood. This PhD research focuses on the detection of changes using TLS (Publications B,
C and D). Sequential datasets were compared to detect 3D temporal variations of the terrain.
Using this method, it is possible to detect sudden changes in the morphology of the slope and
small-scale deformations. Measurement was considerably improved (up to a subcentimetric
scale) by applying a Nearest Neighbour (NN) averaging technique.

The results of this research mainly consisted of: (a) the detection of rockfalls (location,
geometry, magnitude and frequency) that occurred in Castellfollit de la Roca and Puigcercós
during the monitoring period (22 and 10 months, respectively); (b) the simulation, in artificial
scenery, of small-scale deformation prior to failure (slope creep, Terzaghi, 1950) using a
methodology that is able to detect subcentimetric deformation; (c) the detection of this
deformation in natural slopes (Castellfollit and Puigcercós); (d) the spatial prediction of
rockfalls.

The possibility of using TLS to spatially detect pre-failure deformation in natural slopes and to
predict the location of a rockfall, constitutes the main contribution of this research. These results
have significant implications for rockfall risk management, specifically as regards the possible
implementation of the findings in an early warning system.


Keywords: rockfalls; Terrestrial Laser Scanner (TLS); monitoring; change detection; precursory
deformation; spatial prediction; early warning.










RESUMEN
Los desprendimientos de rocas de pequeñas dimensiones (<100m
3
) son el fenómeno
geomorfológico de mayor frecuencia en laderas rocosas, acantilados y/o taludes antrópicos (Guzzetti
et al., 2004; Copons and Vilaplana, 2008; Lim et al., 2009). Estos movimientos pueden alcanzar una
energía de impacto considerable (Agliardi and Crosta, 2003; Dorren and Seijmonsbergen, 2004),
pudiendo afectar a infraestructuras, edificios e incluso personas. El propósito principal de esta
investigación consiste en mejorar nuestro conocimiento actual sobre el fenómeno de los
desprendimientos, empleando para ello la tecnología Light Detection and Ranging (LIDAR) de tipo
terrestre, más específicamente mediante un instrumento denominado Láser scanner Terrestre
(TLS), modelo ILRIS-3D (Optech). Ésta novedosa tecnología permite adquirir las coordenadas
espaciales de millones de puntos (nube de puntos) de la superficie del terreno, con una precisión (σ
= 0.72cm a 50m), resolución (de orden mm-cm) y velocidad (2.500 medidas/seg.) inimaginables
hace tan solo unos años. Durante los últimos años, las aplicaciones del TLS en laderas rocosas ha
aumentado considerablemente, como por ejemplo la obtención de un Modelo de Elevaciones
Digitales (DEM) en zonas escarpadas de alta resolución y precisión, la obtención de la orientación
de los planos de discontinuidad que condicionan los desprendimientos, el monitoreo de laderas, la
obtención de las relaciones magnitud-frecuencia de desprendimientos, etc. (Slob and Hack, 2004;
Bauer et al., 2005; Rosser et al., 2005; Lim et al., 2006; Jaboyedoff et al., 2007; Oppikoffer et al,
2008a; Sturzenegger and Stead, 2009, entre otros). De estas aplicaciones, la posibilidad de emplear
el TLS para predecir, de manera combinada, la localización y fecha de rotura de futuros
desprendimientos supone un gran reto en la actualidad (Rosser et al. 2007; Abellán et al., 2009;
Abellán et al., accepted).

Las áreas de estudio piloto seleccionadas representan una amplia variedad de escenarios
geomorfológicos, litologías, comportamiento geomecánico y niveles de peligrosidad y riesgo. Estas
áreas son: (a) un sector de las abruptas laderas gnéisicas en Vall de Núria (Publicación A); (b) el
acantilado basáltico de Castellfollit de la Roca (Publicación B); (c) el escarpe de coronación, en
materiales margo-calcáreos, del antiguo movimiento de ladera de Puigcercós (Publicación D). El
estudio de estas áreas se ha completado con un ensayo en un escenario artificial, con el objetivo de
realizar un desarrollo metodológico (Publicación C). El énfasis de la investigación ha consistido en
la detección de cambios empleando el TLS (Publicaciones B, C y D). Por ello, la metodología se
basa, principalmente, en realizar un monitoreo de diversas zonas de estudio a lo largo del tiempo. De
este modo, mediante una comparación de las nubes de puntos adquiridas con el TLS de un modo
secuencial, se pueden detectar: (a) cambios bruscos en la morfología de la ladera debido a la
ocurrencia de desprendimientos, y (b) pequeñas deformaciones en la ladera identificadas como
precursoras a la ocurrencia de un desprendimiento (slope creep, Terzaghi, 1950). Esta detección se
mejora considerablemente (hasta un orden subcentimétrico) mediante un filtrado del error
instrumental basado en un promedio de las medidas (NN averaging technique).

Los resultados de esta investigación son principalmente: (a) la detección de las caídas de rocas
(localización, geometría, magnitud y frecuencia) que se han producido en dos áreas de estudio
(Castellfollit de la Roca y Puigcercós) durante un periodo de monitoreo de 22 y 10 meses,
respectivamente; (b) la detección de deformaciones de orden subcentimétrico en un escenario
artificial, simulando el comportamiento que precede a las caídas de rocas; (c) la detección de dicha
deformación precursora en laderas naturales; (d) la predicción espacial de futuros desprendimientos
de rocas.

La posibilidad de emplear el TLS para la detección espacial de la deformación previa a la ocurrencia
de un desprendimiento y, con ello, para la predicción de futuros desprendimientos, constituye la
aportación de mayor alcance de esta investigación. Este resultado tiene importantes implicaciones en
la gestión del riesgo por caídas de rocas, específicamente en lo relativo a su posible implementación
futura en un sistema de alerta temprana (early warning sytem).



Palabras clave: caídas de rocas; Láser Scanner Terrestre (TLS); monitoreo; detección de cambios;
deformación precursora; predicción espacial; detección temprana.







ACKNOWLEDGEMENTS

Al Dr. Joan Manuel Vilaplana, director del grupo RISKNAT y de esta investigación pre-
doctoral. Su supervisión y apoyo constante han permitido que esta tesis fuera posible.
Gracias por confiar en mi durante todos estos años.

I would like to thank Dr. Nicholas J. Rosser, professor at the Institute of Hazard and Risk,
Geography Department, Durham University (UK) for supervising my research during a 3-
month academic visit (2006), for his continuous support of my research, and for passing on
his knowledge of instrument capabilities and cliff monitoring.

Dr. Michel Jaboyedoff, directeur du Institut de géomatique et d'analyse du risque (IGAR) et
professeur de la Faculté des Géosciences et de l'Environnement, Universitéde Lausanne.
Cher professeur, je vous remercie votre supervision de mon stage dans votre Institut (Août -
Décembre 2007). Merci beaucoup aussi pour vos conseils et votre collaboration dans l'étude
des mouvements précurseurs.

Agradecer también al Dr. Jaume Calvet por su apoyo y colaboración en la zona de estudio
piloto de Puigcercós, por sus correcciones de algunas de las partes de esta tesis y por su
buen humor. Al Dr. Josep Anton Muñoz, director del Centre Mixt d'investigació
GEOMODELS por su apoyo a la línea de investigación en la que se enmarca esta tesis:
estudio de caídas de rocas mediante el Láser Escáner Terrestre (TLS). Al Dr. Ramón Copons
por su investigación de desprendimientos de rocas en Andorra la Vella. Su tesis supervisada
también por Dr. JM. Vilaplana ha dejado muy alto el listón para tesis sucesivas. A la Dra.
Emma Suriñach, directora del proyecto DALMASA, por su colaboración en curso (y con
muchas cosas por hacer todavía) en las zonas de estudio del Solá de Andorra la Vella y
Montserrat. Al Dr. Jordi Corominas, por ser un referente nacional e internacional en lo
relativo a los movimientos de ladera. Gracias también a todos los profesores y personal de
administración del Departament de Geodinàmica i Geofísica de la Facultat de Geologia de
la Universitat de Barcelona (en especial a Mª Angels Marqués, Gloria Furdada, Jaume
Bordonau y Josep Maria Casas), por la convivencia durante este periodo de tesis, así como
por facilitarnos la tarea a los jóvenes investigadores del departamento. Dar también las
gracias a mis profesores de la Universidad de Alicante por su apoyo durante mi formación
de grado: Dr. Pepe Delgado, Dr. Salvador Ordóñez, Dr. Juan Carlos Cañaveras y Dra. Mª
Ángeles García del Cura, entre otros.

Quisiera también agradecer a los siguientes organismos e instituciones que han colaborado
en la financiación de parte de la investigación: al Ministerio de Ciencia e Innovación por la
concesión de una beca de Formación de Profesorado Universitario (FPU AP-2004-1852) de
4 años de duración; al Parc Natural de la Zona Volcanica de la Garrotxa (PNZVG) por
financiar el proyecto “Estudi de caigudes de blocs a la cinglera basàltica de Castellfollit de
la Roca”; al Institut GEOMODELS por la concesión de una beca de colaboración de 9
meses para el estudio del fenómeno de desprendimientos de rocas mediante la técnica del
TLS; a los proyectos “Desarrollo de métodos para la detección, caracterización y
prevención de movimientos de masa (DALMASA, MEC CGL2006-06596)”, TopoIberia
CSD2006-0004/Consolider-Ingenio2010 por su apoyo económico en campañas de campo y
asistencia a congresos. Y como no, al Grup de Riscos Naturals (RISKNAT) por acogerme en
un marco institucional en el que desarrollar la tesis.

Agradecer también a George von Knorring por sus correcciones del inglés. A David García
por su colaboración en la adquisición de los datos de campo con el TLS en Castellfollit de la
Roca y Puigcercós. A los compañeros Marc Janeras i Ferran López por su colaboración en
la Vall de Núria y Montserrat. A los investigadores del Institut de Geomática Michele
Crosetto, Oriol Monserrat y Piotr Jaszczak por su colaboración en la fase inicial del estudio
en Castellfollit de la Roca. A los compañeros de mi estancia en Durham (Michael Lim y
Stuart A. Dunning) y Lausanne (Thierry Oppikofer, Andrea Pedrazzini y Julien Travelleti).
Los trabajos de Fin de Carrera y de Fin de Máster llevados a cabo en Castellfollit de la Roca
(Eva Asensio) y Puigcercós (Julien Blanchard; Xavi Rodríguez) me han permitido
profundizar en el conocimiento de la dinámica de los desprendimientos en estas zonas de
estudio. Agradecer también a este último su colaboración dando un cuidadoso formato a
algunos de los capítulos de esta tesis.

Durante estos 5 años de investigación, este trabajo no hubiera sido posible sin la ayuda
diaria de mis amigos y compañeros del departamento (algunos de ellos ya doctores) Dr.
Ignasi Vilajosana, Dra. Marta Guinau, Gia Khazaradze, Dra. Ester Falgàs, Dr. Albeiro
Rendón, Dra. Maria Ortuño, Dr. Ángel Rodés, Micky Marín, Blanca Payás, Xavi Rodríguez,
Carlos Rubí, Dra. Mapi Asta y Clara Torrentó. Muchas gracias a todos. Por último, lo más
importante: gracias a mis padres y hermanos por ayudarme en todo momento. Y, como no,
gracias a Berta por todos estos años compartidos ;b



































INDEX
PhD thesis
















































SECTION I. INTRODUCTION
19

CHAPTER 1. General overview 21


1.1. Structure of the PhD 21
1.2. Justification of the research 23
1.3. Aims 24
1.4. Study areas 24
1.5. Overview of the publications 26




Abstract 31
2.1. Introduction 33
2.1.1. Study area
2.1.2. Terrestrial laser scanner
33
34
2.2. Methodology 35
2.3. Data acquisition 36
2.4. Results: slope and rockfall modelling 37
2.4.1. Inventory of rockfalls 38

2.4.1.1. Coordinates of the detachment area
2.4.1.2. Joint geometry modelling
2.4.1.3. Block geometry and volume
2.4.1.4. Rockfall susceptibility
38
39
40
41
2.4.2. 3D Surface modelling: High accuracy DEM 41

2.4.2.1. Trajectories 42
2.5. Discussion 42
2.6. Conclusions 44

Acknowledgements
References





44
45
SECTION II. CHANGE DETECTION BY TLS
47


CHAPTER 3. Summary of the methodology 50


3.1 Material 51

3.1.1. Instrumental: TLS ILRIS-3D
3.1.2. Data acquisition
51
52
3.2 Methodology 53

3.2.1. Comparison of sequential TLS datasets
3.2.2. Detection of small displacements
3.2.3. Prediction of rockfalls
53
53
53




CHAPTER 4. Rockfall detection.
Case study at Castellfollit de la Roca (publication B)

55
Abstract
57
4.1. Introduction 59

4.1.1. Study area
4.1.2. Historical inventory of rockfalls

60
61

CHAPTER 2. Applications of TLS to rockfall studies.
Case study at Vall de Núria (publication A) 29
4.1.3. Conditioning and triggering factors 61
4.2. Material and methods 63
4.2.1.Terrestrial Laser Scanner (TLS) 63
4.2.2. Data acquisition 63
4.2.3. Comparison of sequential TLS datasets 64
4.2.4. Assessment of the quality of the rock face modelling 65
4.3. Results 66
4.3.1. Detachment of single columns 66
4.3.1. Detachment of groups of columns 66
4.3.3. Slab failures 66
4.4. Discussion
4.4.1. Pros and cons of TLS
4.4.2. Critical review of the results
4.4.3. Combination of long-term and short-term approaches
4.4.4 Implications of the results
68

68
68
68
69
4.5. Conclusions 70

Acknowledgements
References

70
70



CHAPTER 5. Detection of small displacements.
Experimental case study (publication C)

73
Abstract
75
5.1. Introduction 77
5.2. Material and methods 77
5.2.1. Instrument characteristics
5.2.2. Experimental setup
5.2.3. Displacement computation
77
78
79
5.3. Results 79
5.3.1. Displacement based or RAW data 79
5.3.2. Nearest Neighbour averaging 79
5.4. Application to a rockfall event 83
5.5. Discussion and conclusions 84

Acknowledgements
References



87
87

CHAPTER 6. Rockfall detection and prediction.
Case study at Puigcercós (publication D)

89
Abstract
91
6.1. Introduction 93
6.1.1. Study area 93
6.2. Material and methods 94
6.2.1. Terrestrial Laser Scanner (TLS) 94
6.2.2. Data acquisition 95
6.2.3. Comparison of sequential scans 96
6.2.4. Detection of the pre-failure deformation 96
6.3. Results

98

6.3.1. Geomorphological evolution of the rock face
6.3.2. Spatial prediction of rockfall
98
98
6.4. Discussion 98
6.5. Conclusions 103

Acknowledgements
References

103
103
SECTION III. GLOBAL SUMMARY

105

CHAPTER 7. Summary of the results 107


7.1. Main applications of TLS in rockfall studies
110
7.1.1. Data acquisition in a single fieldwork campaign 110
7.1.2. Multi-temporal data acquisition 110
7.2 Rockfall detection 111
7.2.1. Castellfollit de la Roca study area 111
7.2.2. Puigcercós study area 112
7.3 Detection of small displacements 113
7.2.1. Outdoor experiment 113
7.2.1. Application to a rockfall event 115

7.4 Spatial prediction of rockfalls


116
CHAPTER 8. Summary of the discussion 119


8.1. Critical analysis
121
8.1.1. Strengths and weakness of the TLS 121
8.1.2. Comparison with related works 122
8.1.3. Comparison with other techniques 124
8.2. Rockfall detection 125
8.3. Precursory deformation 126

8.4. Spatial prediction of rockfalls


128
CHAPTER 9. Concluding remarks 131


9.1. Summary of the conclusions
132
9.2. Further research 134







APPENDICES 135
Appendix 1: Overall references
137
Appendix 2: List of tables
149
Appendix 3: List of figures
153
Appendix 4: Specific terminology
159
Appendix 5: Extended summary in Spanish
163



























Chapter 1. General Overview


Antonio Abellán Fernández Page 19




I
SECTION

Introduction



SECTION I. Introduction



Page 20 PhD thesis

































Chapter 1. General Overview


Antonio Abellán Fernández Page 21


















CHAPTER 1
General
overview




In accordance with the current regulations for PhDs submitted as a compendium of publications, a
general introductory section and an overview of the different publications is required.


1.1 General introduction

1.1.1 Structure of the PhD

This thesis is divided into four sections and nine chapters, as outlined below:




SECTION I. Introduction



Page 22 PhD thesis

SECTION I. INTRODUCTION
- CHAPTER 1. General overview
- CHAPTER 2. Applications of TLS in rockfall studies (Publication A)

SECTION II. CHANGE DETECTION
- CHAPTER 3. Summary of the methodology
- CHAPTER 4. Rockfall detection (Publication B)
- CHAPTER 5. Detection of small displacements (Publication C)
- CHAPTER 6. Rockfall detection and prediction (Publication D)

SECTION III. GENERAL SUMMARY
- CHAPTER 7. Summary of the results
- CHAPTER 8. Summary of the discussion
- CHAPTER 9. Concluding remarks

APPENDICES


The PhD thesis is introduced in SECTION I. Chapter 1 gives a general overview of the thesis,
including a justification of the study, the aims of the research and an outline of the publications.
Chapter 2 (Publication A) provides an introduction to the application of the terrestrial laser scanner
(TLS) in rockfall studies.

The core of the thesis (SECTION II, CHANGE DETECTION) consists of four chapters (Chapters
3 to 6). Chapter 3 summarizes the methods used in the subsequent chapters. Chapter 4 (Publication
B) deals with a case study of rockfall detection by TLS monitoring at the basalt cliff of Castellfollit
de la Roca. Chapter 5 describes an outdoor experiment performed to detect slow (mm-cm)
displacements (Publication C). The methods developed in this experiment allowed us to identify a
precursory displacement in a 50m
3
rockfall event. Finally, Chapter 6 (Publication D) deals with
rockfall detection and prediction at the Puigcercós test site. Prediction is based on the detection of a
precursory deformation, and represents the main contribution of this PhD to understanding of
rockfalls.

SECTION III provides a GENERAL SUMMARY of the previous sections. Chapters 7, 8 and 9
contain a summary of the results, a summary of the discussion and a summary of the conclusions
respectively, with an emphasis on rockfall detection and prediction. Finally, a list of tables and
figures, a summary of the specific terminology, and an extended summary in Spanish is provided at
the end of the PhD thesis. The final section consists of following APPENDICES: overall references,
list of tables and figures, specific terminology and an extended summary en Spanish.




Chapter 1. General Overview


Antonio Abellán Fernández Page 23
1.2 Justification of the research

Rockfall is the relatively free fall or precipitous movement of a newly detached segment of bedrock
from a cliff or other very steep slope (Bates and Jackson, 1987). Minor scale rockfalls (up to 100 m
3
)
are the most frequent type of landslides on steep slopes, such as rock faces (Copons and Vilaplana,
2008), coastal cliffs (Lim et al., 2009; Dewez et al., 2009) and transportation corridors (Budetta,
2004; Guzzetti et al., 2004). Since rockfall is the fastest type of landslide (Varnes, 1978), its impact
energy (and hence the geological hazard) can be very high (Pfeiffer et al., 1995; Agliardi and Crosta,
2003; Dorren and Seijmonsbergen, 2004). Rockfall presents a common risk to populations,
buildings and/or infrastructures in steep mountainous terrain (Corominas et al., 2003; Copons et al.,
2005). Rockfalls are currently the primary cause of landslide fatalities in some countries (Guzzeti,
2000). When this natural phenomenon interacts with populations, buildings or infrastructures, and
when the value of the risk exceeds that of the acceptable risk (see Fell et al., 2008), the hazard,
exposure or vulnerability parameters should be reduced.

The research deals with minor scale rockfalls, given that the magnitude and frequency of rockfalls
are related by an inverse power law (Douglas, 1980; Teixeira, 2006; Lim et al., 2009), i.e., the lower
the magnitude, the higher the frequency. There are a number of approaches to the study of rockfall
phenomena, such as frequency estimations (e.g. Malamud et al., 2004; Stoffel et al., 2005; Scneuwly
and Stoffel, 2008), susceptibility assessment (e.g. Ayala-Carcedo et al., 2003; Günther et al., 2004;
Frattini et al., 2008), and hazard and/or risk assessment (e.g. Hungr et al., 1999, Copons et al.,
2005; Corominas et al., 2005; Miller et al., 2008). A landslide hazard assessment should ideally
determine when and where a slope failure is likely to occur (e.g. Hartlen and Viberg, 1988; Guzzetti
et al., 2004). As discussed by Van Westen et al. (2006), these two questions do not have a simple
answer. During the last decade, innovative applications in remote sensing techniques have been
developed. The possibility of acquiring datasets with high accuracy and spatial resolution using laser,
optical and/or radar technologies mounted on terrestrial, aerial and/or satellite instrumentals is
currently opening up new ways of visualizing, modelling and interpreting Earth surface processes.

One of these new remote sensing tools, a Terrestrial Laser Scanner (TLS) was used in this study.
This instrument is also known as a ground-based LIDAR (Light Detection and Ranging). The
possibility of studying rockfall phenomenon using this new tool has significant challenged our
research group (Abellán et al., 2006; Blanchard et al., 2008; Vilajosana et al., 2008; Abellán et al.,
2009; Abellán et al., accepted; Guinau et al., accepted; Abellán et al., submitted). This PhD, in
combination with related studies (Rowlands et al., 2003; Bauer et al., 2005; Rosser et al., 2005;
Biasion et al., 2005; Rosser et al., 2007; Armesto et al., 2009; Oppikofer et al., 2009; Dewez et al.,
2009; Lim et al., 2009), clearly demonstrated the potential of TLS technology in rockfall studies.
The findings of this research are of paramount importance for the early detection of rockfalls, which
may constitute a step forward in rockfall risk management.





SECTION I. Introduction



Page 24 PhD thesis
1.3 Aims

The overall aim of this PhD was to improve our understanding of the rockfall phenomenon by
means of the new TLS technology.

The thesis also had the following specific aims. The name of the publication in which the specific
aim is discussed is shown in brackets.
(i) To improve the geometric characterization of rock slopes (Publication A).
(ii) To better understand the geomorphological evolution of rock slopes through
sequential acquisition of TLS datasets (monitoring), including the calculation of the
volume and frequency of rockfalls during the survey period (Publications B and D).
(iii) To develop a method for detecting centimetric and/or sub-centimetric deformations
(Publication C).
(iv) To apply this method to detecting precursory deformation in rockfalls that took
place during the surveying period (Publication C).
(v) To apply this method to the spatial prediction of rockfalls (Publication D).

The research is focused on the use of TLS to detect change (specific aims ii, iii and iv) and its
application to the spatial prediction of rockfalls (specific aim v). The potential of TLS to detect
unstable areas prior to the occurrence of a slope movement is one of the most innovative aspects of
this research. The findings have a great significance from a scientific and an applied perspective,
specifically in terms of the potential use of TLS in rockfall early warning systems that cover wide
areas.



1.4 Study areas

The selected study areas are characterized by highly variable lithology (gneisses in Vall de Núria;
basalts in Castellfollit de la Roca; and marls and sandstones in Puigcercós), geomorphological
landscapes (a high mountain area, a cliff eroded by a river and scarp formed by an old landslide),
geomechanical behaviour (brittle vs. ductile), failure mechanisms (slide, fall and toppling), hazard
levels (different magnitude-frequency relationships), the existence of vulnerable elements (a rack
railway, houses at the edge of a cliff or no exposed elements) and risk. A summary of these
characteristics can be found in Table 1.1.







Chapter 1. General Overview


Antonio Abellán Fernández Page 25



VALL DE NÚRIA
(Chapter 2)
CASTELLFOLLIT DE LA
ROCA (Chapter 4)
PUIGCERCÓS
(Chapter 6)
Lithology
Gneisses and schists. Multiple
families of foliation and
diaclasis were found.
Two lava flows linked by an irregular
deposit formed by pyroclasts and
paleosoils. Quaternary fluvial
deposits at the bottom of the
sequence.
Alternation of grey marls,
sandstones, silt and clays (sub-
horizontal bedding).
Geomor-
phological
landscape
Steep rocky slopes. High
mountain valley modelled by
fluvial and glacial processes.

The study area is located in the
Garrotxa volcanic field. The lava
flows were laterally eroded by the
Fluviá River to form a cliff.
(dimensions =8000 m
2
; maximum
height: 40m).
The pilot area corresponds to the
main scarp of a landslide that
occurred in 1881. Dimensions:
3000m
2
(max. height =25m).
Geomecha-
nical beha-
viour and
failure
mechanism

Brittle failure. Stability
controlled by fault and joint
families. The intersection of
these discontinuities generates
wedges. Failure mechanism =
slide.
The pyroclasts and paleosoil level
was identified as a weakness. The
pre-existing columnar jointing pattern
plays an important role in the
geometry of the detached blocks.
Two failure mechanisms were found:
fall and toppling. A precursory
deformation was observed prior to
the occurrence of a 50m
3
rockfall.
An alternation of ductile and brittle
materials (marls and sandstones,
respectively) was found.
Extensional discontinuities due to
the 1881 landslide control the
current rockfalls. Failure
mechanism: toppling. Precursory
deformation was detected prior to
the occurrence of different
rockfalls.
Hazard
(volumes,
frequencies
and
velocities)
Most frequent volumes range
from 1 to 10 m
3
. Mean rockfall
annual frequency in the valley:
5-6 events (see Fernandez and
Vilaplana, 2004). Considerable
hazard due to steep slopes
and high fall height. As a
result, high energy and
bounces and long runout
distances are characteristics of
this area.
The most frequent volumes range
from 0.5 to 1.5 m
3
. The annual
rockfall frequency (4 rockfalls/ year)
is considerably lower than that of the
Puigcercós study area (see Chapter
6). Free fall height ranges from 30 to
40 meters. Short runout distances.
Two rock slab failures (magnitudes
around 1000m
3
) occurred in the last
30 years.
The magnitude of the most
frequent rockfalls is below 0.5 m
3
.
A very high rockfall frequency was
found (more than 50 rockfalls with
magnitudes >0.1 m
3
occurred in
2008). Free fall height ranges from
10 to 25 m. Short runout
distances.
Exposure
(existence of
vulnerable
elements)
Two vulnerable elements: the
rack railway and walkers on
the Queralbs-Núria path.

There are two vulnerable scenarios:
the top (Castellfollit de la Roca
Village) and the bottom of the cliff. No
permanent vulnerable elements were
found at the latter.
There are no permanent
vulnerable elements in the study
area. The old village of Puigcercós
was abandoned prior to the
occurrence of the 1881 landslide.
Risk
A recently constructed tunnel
means that the train is no
longer exposed to rockfall.
However, walkers face a
certain degree of risk.
The considerable cliff retreat rate
may endanger the houses located on
the edge. The risk is lower for the
bottom of the cliff.
Reduced risk due to low exposure.

Table 1.1. Summary of the main characteristics of the different pilot study areas. A detailed analysis (except for
risk and exposure assessments) is discussed in subsequent sections (see Chapters 2, 4 and 6).







SECTION I. Introduction



Page 26 PhD thesis
1.5 Overview of the publications

The number of publications discussing the use of TLS in the Earth Sciences field has grown
considerably during the last five years (2003-2008). As an example, Figure 1.1 shows the number of
publications in the aforementioned field found in the “Science Citation Index Expanded (SCI)”.


0
5
10
15
20
25
30
35
40
2000 2001 2002 2003 2004 2005 2006 2007 2008
year
n
u
m
.

o
f

p
u
b
l
i
c
a
t
i
o
n
s
laser+scan*+terrest*
laser +rockfall

Fig. 1.1. Number of publications indexed by the SCI Expanded (ISI Web of Knowledge,
www.isiwebofknowledge.com). Analysed period from 2000 to 2008. Keywords (in topic) =(laser scan* terrest*) or
(laser rockfall); Subject Areas =geography or geology; Current year (2009) is not included in this figure due to an
incomplete record.


An overview of the four PhD thesis publications is given below.

PUBLICATION A
1
(see Chapter 2) deals with the main application of the terrestrial laser scanner
to rockfall studies. The test area (Vall de Núria, Eastern Pyrenees, Spain) was selected as it is the site
of frequent rockfall activity (Rendón, 2004; Fernández and Vilaplana, 2004) that affects the rack
railway. Two rockfalls of dozens of cubic meters each damaged the rack railway and isolated the
Vall the Núria resort in April and J une 2003. This highlights the rockfall hazard and risk in the area.
To improve our understanding of rockfall hazard in the Vall de Núria, we needed to study the source
and propagation areas. At that time, the TLS attracted our attention, as this innovative technology
could acquire dense 3D information with centimetric accuracy. Although literature on the
application of TLS to natural hazards was scarce (e.g. Hunter et al., 2003; Rowlands et al., 2003),
we launched a TLS field study in a section of this valley. The dense point cloud obtained with this
technology was the basis for the following applications: (a) a high accuracy & resolution digital
elevation model (DEM) that considerably improved the subsequent rockfall simulation (trajectory

1
Abellán, A., Vilaplana, J. M., and Martínez, J.: Application of a long-range Terrestrial Laser Scanner to a
detailed rockfall study at Vall de Núria (Eastern Pyrenees, Spain), Eng. Geol, 88(3–4), 136–148, 2006



Chapter 1. General Overview


Antonio Abellán Fernández Page 27
and energy); (b) accurate 3D location, volume calculation and orientation of the family of
discontinuities that conditioned one of the rockfall events mentioned above.

This first publication was of interest to my PhD as it opened up a wide range of potential
applications of the instrument. To deepen my research, I had to select and develop just one
application. Finally, the following topic was selected: rockfall monitoring (detection and prediction)
through a multi-temporal comparison of TLS datasets. At the end of 2005, innovative research on
the application of TLS to monitoring a cliff section on the coast of North Yorkshire (UK) was
published (Rosser et al., 2005). Monthly acquisition of datasets enabled the Durham University
group to calculate the volume and the rockfall frequency with a previously unobtainable precision. I
started working on this topic by acquiring datasets on the emblematic basalt cliff at Castellfollit de la
Roca (Natural Park of the Garrotxa Volcanic Field, Catalonia, Spain). A few months later I had the
chance to apply for an academic visit to Durham University. Consequently, for 3 months I was
supervised by Dr. Nicholas J . Rosser, the author of the aforementioned publication (Rosser et al.,
2005). The method developed by this group, i.e., change detection using TLS, was the basis for the
further development of this PhD thesis.

PUBLICATION B
2
(see Chapter 4) shows the results of monitoring the basalt cliff at Castellfollit
de la Rocafor 22 months (from March 2006 to J anuary 2008). The vulnerability of the houses at the
top of the cliff was the main reason for carrying out this research. Two types of mass movement
were detected in the monitoring period, which enabled us to make a short-term estimate of the
frequency of minor scale rockfalls: (a) detachment of single basalt columns, with magnitudes below
1.5 m
3
and (b) detachment of group of columns, with magnitudes from 1.5 to 150 m
3
. These results
were temporally biased: a longer study period was required. As a result, this approach was combined
with historical records to obtain a more representative estimation of the rate of retreat of the cliff.

Undoubtedly, the most important finding in this study area was the detection of a centimetric
deformation prior to a 50m
3
rockfall (April 2007 event). This deformation was interpreted as a slope
creep phenomena (Terzaghi, 1950), i.e. a precursory deformation prior to the occurrence of a mass
movement. Nevertheless, this centimetric movement was disguised by same-order-of-magnitude
instrumental noise. Thus, a filtering technique was needed. At this time, there were no publications
in the literature on the use of TLS to detect precursory deformations. Therefore, the prediction of the
event was impossible technologically and conceptually.



2
Abellán, A., Vilaplana, J.M., Calvet, J., García-Sellés, D., Asensio, E., submitted. Rockfall monitoring by
Terrestrial Laser Scanning. Case study of the basaltic rock face at Castellfollit de la Roca (Catalonia, Spain). Nat.
Hazards Earth Syst. Sci.




SECTION I. Introduction



Page 28 PhD thesis
PUBLICATION C
3
(see Chapter 5) was carried out in collaboration with the Institute of
Geomatics and Risk Analysis (IGAR), University of Lausanne (Switzerland), and was supervised by
Dr. Michel Jaboyedoff. An outdoor experiment was performed to ascertain whether the TLS
instrumental error was small enough to detect precursory, small scale (mm-cm) deformation. The
experiment consisted of an induced displacement of three objects relative to a stable part.
Millimetric changes were not detected using conventional methods. However, the accuracy of the
measurements was considerably improved through an averaging technique (25 Nearest Neighbours).
This study allowed us to accurately detect the location and magnitude of small displacements from a
stable part, which indicated that precursory displacements on real rock slopes could also be detected.
Subsequently, the technique was applied to the precursory deformation observed in the Castellfollit
de la Roca datasets (April 2007 event). This deformation was clearly detected. The implication of
these results very important: if millimetric displacements prior to a rockfall can be detected through
a TLS, the spatial prediction of rockfalls could be a real possibility. Nevertheless, in this case, the
precursory deformation was detected through a back analysis procedure. Thus, it was not a real
prediction. At this stage of the PhD thesis, the detection of precursory deformation in additional
study areas was necessary.

Finally, PUBLICATION D
4
(see Chapter 6) shows the application of the TLS to rockfall detection
and prediction at the main scarp of Puigcercós landslide. The first part of the research was carried
out using the same approach as that described in publication B: change detection by TLS. As a result,
we detected more than 40 rockfalls (from 0.1 to 100 m
3
) in a 9 month period (from October 2007 to
J uly 2008). This was a much higher rockfall frequency than previously expected. The second and
most relevant part of this publication dealt with detecting a centimetric precursory deformation in
different parts of the slope. In contrast to the back analysis procedure used in publication B, the early
detection of a precursory deformation was the basis for the spatial prediction of rockfalls.

The research carried out in the Puigcercós test area allowed us to deepen our understanding of
precursory deformation in minor scale rockfalls. The possibility of linking spatial and temporal
prediction through the detection of precursory deformation could be of paramount importance to
rockfall early warning systems. Nevertheless, more case studies are needed of different materials,
structural conditions and different failure mechanisms.







3
Abellán, A., Oppikofer, T., Jaboyedoff, M., Vilaplana, J.M., 2009. Detection of millimetric deformation using a
terrestrial LIDAR: Experiment and its application to a rockfall event. Nat. Hazards Earth Syst. Sci., 9, 365–372,
2009

4
Abellán, A., Vilaplana, J. M., Calvet, J., Blanchard, J., 2009_accepted. Detection and spatial prediction of
rockfalls by means of terrestrial laser scanning monitoring. Geomorphology.





























CHAPTER 2
Applications of TLS to
rockfall studies


Publication A: Abellán et al., (2006). Engineering Geology
































































Application of a long-range Terrestrial Laser Scanner
to a detailed rockfall study at Vall de Núria (Eastern
Pyrenees, Spain)

A. Abellán
a,*
, J.M. Vilaplana
a,
, J. Martínez
b

a
RISKNAT group & GEOMODELS Institute, Depart. of Geodynamics and Geophysics, Univ. of Barcelona, Spain
b
Serfocar S.L.: Servicios de Fotogrametría y Cartografía. C/ Saragossa, 95-97. 08006-Barcelona, Spain


Submitted: December 2004 - Revised: December 2005 - Accepted: September 2006 – Published: December 2006

Abstract. In this study we show the application of a long-range Terrestrial Laser Scanner (TLS) to a
detailed rockfall study in a test zone at Vall de Núria, located in the Eastern Pyrenees. Data
acquisition was carried out using TLS-Ilris3D, the new generation of reflector-less laser scanners with
a high range, accuracy and velocity of measurements. Eight scans were performed at 3 stations to
acquire coordinates of almost 4 million points. The results from the acquired data are a high accuracy
Digital Elevation Model (DEM) and the reconstruction of the joint geometry. The former is used for
inventory of rockfalls and for more accurate rockfall simulation (trajectories and velocities). The latter
allows us to model the geometry and volume of the source area in recent rockfalls. Our findings
suggest that TLS technology could be a tool of reference in rockfall studies in the near future.

© 2006 Elsevier B.V. All rights reserved.


Keywords: Rockfall; Terrestrial Laser Scanner; DEM; Rock slope; 3D geometry modelling; Eastern
Pyrenees















Abellán et al. (2006) Chapter 2: Applications of TLS to rockfall studies





Antonio Abellán Fernández Page 33

2.1 INTRODUCTION

Rockfall is the relative free falling or precipitous
movement of a newly detached segment of bedrock
from a cliff or other very steep slope (Bates and
Jackson, 1987). Since rockfall is the fastest type of
landslide (Varnes, 1984), it presents a common risk
to transportation and structures in steep
mountainous terrain (Pfeiffer et al., 1995;
Corominas et al., 2003).

In recent years considerable advances have been
made in the analysis of rockfall susceptibility
(Marquínez et al., 2003; Baillifard et al., 2003;
Günther et al., 2004), in the modelling of rock
trajectories on a 3D slope (Agliardi and Crosta,
2003; Crosta and Agliardi, 2004; Dorren and
Seijmonsbergen, 2004), as well as in the risk
management of rockfalls (Guzzetti et al., 2004;
Copons et al., 2005). Many of these studies use
topographical maps and Digital Elevation Models
(DEMs) derived from aerial sensors, i.e. aerial
photography and airborne LIDAR (Light detection
and Ranging). These sensors achieve maximum
density of information when the incident ray is
perpendicular to topography, typically subhorizon-
tal surfaces. By contrast, the instabilities due to
rockfall usually occur on vertical slopes with the
result that a greater density of information using
terrestrial sensors is obtained.

One of the terrestrial sensors is the new long-range
Terrestrial Laser Scanner (TLS). This is a technolo-
gy with considerable potential in the characteriza-
tion and monitoring of slope instability due its
maximum reach (10
3
m) and accuracy (7mm at 50m).

TLS was used at Vall de Núria (Eastern Pyrenees),
an area currently affected by rockfalls (Rendón and
Vilaplana, 2004). In our study we show how TLS
can help us to improve the estimation of parameters
of interest in rockfall studies: TLS can produce high
resolution DEMs, which can be employed for
inventory of rockfalls, monitoring of mass
movement time evolution and more accurate
numerical simulation of rockfall trajectories and
velocities. Furthermore, data obtained from TLS
allows the reconstruction of joint geometry and an
estimation of the volume of blocks that can fall
from steep inaccessible rock slopes.


2.1.1 Study area
The study area is located in a high mountain valley,
Vall de Núria, Eastern Pyrenees, Spain (Fig. 2.1).
There is only one form of transport in this area: the
Fig. 2.1. Study area in Vall de Núria, Eastern Pyrenees, Spain. Pilot area is enclosed by dotted line.





SECTION I. Introduction Publication A: Eng. Geol., 88(3–4), 136–148



Page 34 PhD thesis

rack railway. Vall de Núria is a popular mountain
resort, especially with skiers and ramblers. It caters
for more than 250,000 visitors annually.

A small test area was selected on the railway track
(300×500 m) in order to calculate the 3D geometry
of the slope using TLS. This area forms part of a
steep rock slope made up of heavily fractured
Paleozoic gneisses of Núria (Santanach, 1974). This
area was selected given its exposure to frequent
rockfalls, resulting in considerable damage to the
railway in recent years (Rendón and Vilaplana,
2004) (Fig. 2.2), specifically in 2003: 3 March, 4
April and 15 June (henceforth, events A, B and C
respectively). This scenario of natural risk has been
the subject of recent studies in an attempt to
evaluate the phenomenon of rockfall in the whole
valley (Rendón et al., 2004; Fernández and
Vilaplana, 2004) and to implement preventive and
corrective measures.


1.2 Terrestrial laser scanner
The long-range TLS is a new instrument that
massively captures coordinates of ground points in
3D with high velocity and accuracy. TLS began to
be used in the 1990s for mobile robot navigation
(Singh and West, 1991; Hancock et al., 1998), in
the construction of metric scale 3D models, such as
sculptures (Beraldin et al., 2000) and industrial
applications (Sequeira et al., 2003). Given the rapid
development of technology, the maximum distance
of the laser is continually being improved. This has
been accompanied by an increase in TLS
applications, including calculation of 3D models of
large surfaces (architecture, archaeology,
topography) and recently the characterization and
monitoring of natural hazards, e.g. volcanoes
(Hunter et al., 2003) and landslides (Rowlands et al.,
2003; Bitelli et al., 2004).

TLS consists of an instrument for measuring
distance (laser) and a scanner. The laser beam is
focused and is reflected directly on the land surface,
obviating the need for the existence of intermediate
prism reflectors. The TLS used in this work is the
Ilris3D from OPTECH (Fig. 2.3). It can reach an
accuracy of 7mm at a distance of 50 m, which is
comparable to a reflectorless total station. TLS is
able to acquire a mean of 2000 points/s (see the
website http://www.Ilris-3D.com). This rapid data
acquisition enables us to acquire 3D coordinates of
millions of points in only 15 min. These points can
be transferred to CAD software (Computer Aided
Design) for subsequent visualization and data
treatment.

After scanning in the study area, the instrument
stores data in a memory card. The file is structured
Fig. 2.2. Rack railway and recent rockfalls in the study area. (a) Event B: 2003/04/04. (b) Event C: 2003/06/15





Abellán et al. (2006) Chapter 2: Applications of TLS to rockfall studies





Antonio Abellán Fernández Page 35

in three columns with the coordinates (x, y, z) of the
total number of points scanned, and a fourth column
corresponding to the value of intensity (l) of the
land surface on which the laser beam is focused.
The maximum range depends on the reflectivity of
the materials on the slope and on the angle of
incidence. Even though range of 1500 m can be
theoretically reached for targets with reflectivity of
80%, the maximum range for natural slopes is in the
order of 1000mowing to the low reflectivity of the
ground. Table 2.1 shows the technical features of
TLS-Ilris3D. A comparison between the existing
TLSs on the market (manufacturers: Cyra, Mensi,
Optech, Riegl, Z and F) can be found in Staiger
(2003).

TLS-Ilris3D is equipped with a LCD (Liquid
Crystal Display), where the zone to be scanned is
visualized (Fig. 2.3a). The user can define the
necessary parameters for data acquisition (spacing
between points, dimensions of the area to be
scanned, etc.) in a PDA (Personal Digital Assistant)
connected to the instrument.



2.2 METHODOLOGY

Fig. 2.4 illustrates the methodology used for the
application of TLS in our rockfall study at Vall de
Núria. This consists of the following steps: (a) data
acquisition: this was carried out using TLS-ILRIS
3D with 4 million 3D points (x, y, z); (b) data
treatment: this was undertaken by triangulation
(Delaunay), trigonometrical calculations, interpo-
lation (Kriging) and rockfall simulations; (c)
calculation of results: high resolution DEM,
trajectories and energies, inventory of rockfalls, Dip
and Dip Direction of joints and volume of the
blocks in the rock slope; (d) validation of results:
this was performed with direct measurements of the
joint orientation in the field by means of a compass,
Fig. 2.3. Close up (3a) and front view (3b, from www.ilris-3D.com) of long range Terrestrial Laser Scanner Ilris-3D

Characteristic Value
Performance
Range

Data sample rate

350 m (4% target)
800 m (20% target)
2,000 points/second
Accuracy
Target registration accuracy
Modeling accuracy
Depth resolution

4 mm
3 mm
3 mm
Size
Scanner (L × W × H)
Scanner weight

312 × 312 × 205 mm
12 kg
Output Metafile consisting of XYZ,
Intensity, digital photo data,
operator setup parameters
and notes.
Spot Spacing



S = 0.026R, where
S = spacing (mm)
R = range to target (m)

Laser Spot Size



D = 0.17R+12, where
D = diameter of spot (mm)
R = range to target (m)
(e.g. 29 mm at 100 m)

Table 2.1: Terrestrial Laser Scanner Ilris3D technical
specifications (www.ilris_3D.com)




SECTION I. Introduction Publication A: Eng. Geol., 88(3–4), 136–148



Page 36 PhD thesis

and through direct observations of trajectories,
volume of the blocks and location of the detachment
area of rockfalls occurred during 2003 (Events A, B
and C). The results that were not validated were
revised in step 2 (data treatment).

The ultimate aim of the study is hazard assessment
(Step 5). This step is currently under construction
given that event frequency has not yet been
estimated.



2.3 DATA ACQUISITION

A minimum of 3 stations: E
1
, E
2
and E
3
were
necessary to cover the study area. From these
stations, scanning with TLS-Ilris3D was performed.
In order to obtain a complete dataset for the study
area, 8 scans were performed as follow: E
1
(E
1
1
,
E
1
2
), E
2
(E
2
1
, E
2
2
, E
2
3
, E
2
4
) and E
3
(E
3
1
, E
3
2
).. The
views from each of the stations and to the scanning
areas are shown in Fig. 2.5.

The total scan time from the sum of the 8 scans was
less than 2 h. One working day was employed for
mobilization, setting up and acquisition of nearly 4
million points using TLS. If a total station had been
used to carry out the same work, at least two weeks
would have been necessary. Moreover, the point
cloud obtained in this way would have been much
smaller. The data acquisition is summarized in
Table 2.2. After selecting the area to be scanned, the
TLS performs a preliminary reading of the mean
distance to the surface of the slope with the aim of
identifying different point spacing. Subsequently,
one of these point spacing is selected and
acquisition of coordinates is carried out. In
accordance with Alba et al. (2005), one of the main
problems in laser scanning data processing is
dealing with huge point clouds. For this reason, a
balance must be reached between information
density acquired with TLS and information that is
necessary. In our study we selected point spacing
from 6 to 20 cm.: E
1
1
: 10cm, E
1
2
: 12cm, E
2
1
: 15cm,
E
2
2
: 13cm, E
2
3
: 20cm, E
2
4
: 6cm, E
3
1
: 20cm, E
3
2
:
20cm. The scan with the highest density of points
(E
2
4
) corresponds to the detachment area of event C,
Fig. 2.4. Methodology used in our study. Step 5 is currently under construction (see section 2.2).





Abellán et al. (2006) Chapter 2: Applications of TLS to rockfall studies





Antonio Abellán Fernández Page 37

which was studied in greater detail (6 cm). Point
spacing from 10 to 20 cm. was good enough to
generate a high accuracy DEM for rockfall
simulation.

We used optimized software (polyworks,
http://www. innovmetric.com) to handle large data
sets obtained from the laser scanner. With this
program, the union of the different scans was
carried out using a user supervised identification
algorithm of the common areas. Owing to our
limited experience in data acquisition using TLS,
the scans E
1
2
and E
3
2
were unsuccessful because of
instrument errors with the result that the
overlapping area was not sufficiently large to join
all the scans. For this reason, the georeferencing of
those scans was carried out by the identification of
control points in the point cloud and in the
photographs and the subsequent acquisition of its
coordinates with a reflectorless total station (Leica
TRC 705).

In the subsequent visualization of the complete
scanned data, it is possible to observe areas without
3D land information (shaded areas). In order to
avoid the appearance of these shaded areas,
additional scans from different angles and heights
should be performed or aerial and terrestrial laser
techniques should be combined (Ruiz et al., 2004).



2.4 RESULTS

slope and rockfall modelling From 4 millions 3D
points we obtained on the one hand the superficial
geometry of the slope (DEM), and on the other hand
the geometry of joints. Our initial hypothesis was
that the bigger and the more accurate the 3D point
cloud, the more detailed the topography of the slope.
Fig. 2.5. Digital Elevation Model (a) and photographs of the study area seen from the stations E2 (b), E1(c) and E3(d). A total of 8
scans were performed from each of the stations: E1 (E11, E12), E2 (E21, E22, E23, E24) and E3 (E31, E32). The scanned areas
are enclosed by dotted lines.







SECTION I. Introduction Publication A: Eng. Geol., 88(3–4), 136–148



Page 38 PhD thesis

Likewise, the more detailed the topography (defined
by DEM) and the more precise the delimitation of
the source area of the rock blocks, the more
accurate the rockfall modelling.

2.4.1 Inventory of rockfalls
Using CAD software allows us to visualize the
slope in 3D, rotate the axis to obtain a better
visualization and to make a zoom or move away
from a specific detail (Fig. 2.6). This visualization
from 3D point cloud or DEM renders geological
and geomorphological features more accurately
than in a topographic map. For example, we can
locate in the point cloud the rockfall source
(bedrock outcrops, potentially unstable rocks,
indicators of recent rockfall, etc.), the rockfall
trajectory and all the relevant information for the
subsequent rockfall simulation.

2.4.1.1 Coordinates of the detachment area
Fig. 2.7 shows the precise location of the source
area of event A at the northern outlet of the Fenech
tunnel. The range of coordinates of this area in the
local system of TLS (8370 m<X<8376 m; 4211
m<Y<4213 m; 1618 m<Z<1620 m) was obtained
from the point cloud. Given that the geometry of the
slope and height of the source area were obtained

with greater accuracy using the point cloud than
with a conventional topographic map (1:5000), the
back analysis of this event was more exact than the
analysis previously carried out by our group
(Rendón, 2004). In this earlier study, the scar was
located with an error of 0.5 cm on the map with
respect to its real position. Now with high accuracy
DEM we can see how high elevation errors (15 m)

Characteristic Value
Total points 3.979.436
Scans 8
Stations 3
Maximum range 870 m
Minimum range 180 m
Accuracy 3 mm @100m

Table 2.2. Data acquisition using TLS-Ilris3D.

Fig. 2.6. (a) Point cloud acquired with TLS Ilris-3D. Selected area is enlarged in Fig. 2.6c. (b) Photography from station E3.
Fig. 2.6c is enclosed with discontinuous line. (c) and (d) different views of the point cloud. Selected area is enlarged in Fig.
2.6d






Abellán et al. (2006) Chapter 2: Applications of TLS to rockfall studies





Antonio Abellán Fernández Page 39

can derive from even small location errors (0.5 cm)
in areas with steep slopes. This variation in the
location of the detached area (from 30 to 45 m
height) implies a variation in the velocity of the
falling rock at the base of the cliff of 5.45 m/s
(Dorren, 2003, Eq. (1))

Velocity = (2 * gravity * fall height)1/2, (eq. 1)

Given that the Energy is 1/2 mass · velocity2 this
change in velocity implies a high variation in
impact energy (from 294 to 442 Kj. for a rock of 1 t
of mass). Likewise, the calculation of the
coordinates of the detachment area using the
geomorphological indicators obtained by Rendón
and Vilaplana (2004) was estimated using the high
accuracy DEM. Thus, the errors associated with the
transfer of information from an oblique photograph
of the slope to the topographic map in 2D were
minimized.


2.4.1.2 Joint geometry modelling
Rock masses can be viewed as being composed of
blocks of intact rocks and joints or discontinuities.
Based on the assumption that rockfall is produced
by the joints in the bedrock (Rouiller et al., 1998), it
is important to characterize these joints to be able to
determine block volume and rockfall susceptibility
in different areas of the rock slope. In rock
engineering, a non-reflector total station (Feng et al.,
2001) and a cyrax laser scanner (Slob et al., 2002)
have been used to measure fracture orientation at
exposed rock faces. Given the small extension of
outcrops in these studies (in the order of 10–100
m2), the ranges were metric or decametric, the
optimal effective range of the scanner being less
than 100m(http:// www.zf-laser.com). However, in
our case, the study area has an extension in the
order of 150,000 m2. So data acquisition with a new
generation of long-range TLS is needed to reduce
the number of scans.

The orientation of a plane can be defined using 3
points, as shown in many handbooks (e.g.
Bronshtein et al., 2003; Slob and Hack, 2004).
Following Feng et al. (2001), if instead of three
points, we have n points obtained by a laser scanner,
we can define the orientation of the best-fit plane
through a planar regression. Accordingly, we
applied a macro (Fernández, 2005) performed in a
commercial software (Microstation, Bentley). As a
result we obtained the Dip Direction (DD) and Dip
Angle (DA) of different joints conditioning
rockfalls that affect the rack railway. In some parts
of the study area, the high altitude scarp did not
allow us to perform manual measurements using the
classic techniques (compass).


Fig. 2.7. Photography (left) and identification of the detachment area in the point cloud acquired with the TLS-Ilris3D (right).







SECTION I. Introduction Publication A: Eng. Geol., 88(3–4), 136–148



Page 40 PhD thesis

The post processing used to calculate joint
orientations was semi-automatic, i.e. the points that
belong to the same joint were selected with the aid
of a visual criterion and the tool ‘Fence’ of
Microstation, (see Fig. 2.8). After this selection, the
macro is operated, yielding the orientation of the
joints. The results obtained for the three joints that
characterize the source zone of event C are as
follows (DD/DA): J1: 143/54; J2: 239/87; J3: 51/52.
The data were validated with direct measurements
of the orientation of the joints.
2.1.3 Block geometry and volume
In agreement with Barton and Bandis (1990), the
block volume parameter plays an important role in
geomechanical behaviour of the bedrock. It is one
of the factors that control the energy of the process
(the others being rock density and fall height). The
block size can be defined using the distance and the
angles between joints (Palmström, 1995). As an
example, we used the demo version of 3D software
joints (http://www.rock-mechanics-software.com)
to calculate the block volume using the orientation
Fig. 2.8. (a) 3D point cloud of the detachment area of event C (2003/06/15). Selected area is enlarged in fig. 8b. (b). 3D point
cloud allows us to model joint geometry: J1, J2 and J3. (c) Photography of the same area. (d) Stereographic projection of the
joints: J1, J2, J3 and direction of movement of the wedge between J1 and J2.






Abellán et al. (2006) Chapter 2: Applications of TLS to rockfall studies





Antonio Abellán Fernández Page 41

and position of a series of joints. Both parameters
(orientation and position) were defined using the
point cloud acquired with the TLS. The results of
the geometry and volume of event C are shown in
Fig. 2.9. The volume calculated for the polyhedron
detached in the event C is 26.53 m3.

2.4.1.4 Rockfall susceptibility
In line with Bieniawski (1993) and Rouiller et al.
(1998), joints are weak areas where rockfalls occur.
The relative orientation of joints with respect to the
slope surface is a fundamental factor in rockfall
susceptibility analysis (Romana, 1993; Günter,
2003). Knowledge of the geometrical characteristics
(DD and DA) of these joints and their relative
position with respect to slope orientation and
inclination would enable us to map areas
susceptible to rockfalls.

Fig. 2.8(d) shows a stereographic projection of the
joints (J1, J2 and J3) that define the detachment area
of event C. As observed in this figure, the wedge
sliding between J1 and J2 is free to slide. We are
currently studying the possibility of determining
rockfall susceptibility as a function of the
geometrical parameters for all the possible
polyhedron existing in a 3D point cloud acquired
with TLS.


2.4.2. 3D Surface modelling: High
accuracy DEM.

In accordance with Hoek (2000), the slope is one of
most important parameters in rockfall simulation. In
the early studies of rockfall simulation, the slope
was represented using a bidimensional profile
(Piteau and Associated Limited, 1980). However, at
present there are numerous rockfall studies where
the topography is described in a 3D mode using a
DEM (Copons et al., 2000; Agliardi and Crosta,
2003; Dorren and Seijmonsbergen, 2004). The
presence of 3D variations in slope morphology
(ridges, convex talus cones, micro-topography)
exerts a considerable influence on rockfall
trajectories (Giani et al., 2004; Crosta and Agliardi,
2004), which are an important factor in rockfall risk
assessment (Copons et al., 2005). Accordingly, the
morphology of the slope defined by a DEM should
be obtained with maximum accuracy. The
characteristics of DEM obtained in our study are as
follows: 1.56·105 m2, 4346 square cells of 6 m
mesh dimension (Table 2.3).
Fig. 2.9. Vertexes coordinate and planes orientation of the detached block of event C. DD: dip direction; DA: dip angle







SECTION I. Introduction Publication A: Eng. Geol., 88(3–4), 136–148



Page 42 PhD thesis

Characteristic Value
Type of interpolation Kriging
Type of mesh Square
Mesh dimension 6 meters
X range (meters) 3949 < X < 4267
Y range (meters) 7974 < Y < 8466
Z range (meters) 1260 < Z < 1925
Total cells 4346
Table 2.3: Characteristics of Digital Elevation Model
generated using the point cloud acquired with the TLS-
Ilris3D


2.4.2.1. Trajectories
We assume that if we obtain a DEM generated with
very accurate data, the results of the rockfall
simulation (trajectories, energy and rebound height)
will bear a closer resemblance to reality. This
assumption was verified employing a deterministic
model called Rotomap (Scioldo, 1991), a three-
dimensional code used for the determination of the
bounce height, energy and path of falling boulders.
With this code, we carried out a rockfall simulation
using two DEMs of different accuracy: DEM 1,
generated with the point cloud obtained with the
TLS (Fig. 2.10a); DEM 2, generated using the
digital topographic map of the Institut Cartogràfic
de Catalunya at scale 1:5000 with lower accuracy
than DEM 1 (Fig. 2.10b). In order to compare
rockfall simulations and to avoid major problems
due to the influence of a detailed topography on
rock fall modelling such as lateral dispersion
(Crosta and Agliardi, 2004), we generated both
DEMs with a similar mesh size: 6 m. Although the
mesh size was constant, contour plots of DEM2 are
smoother than those of DEM 1 (Fig. 2.10). This
effect is more pronounced in areas with steep slopes
than in areas with gentle slopes. Given the accuracy
achieved with TLS, DEM 1 is more reliable than
DEM 2.

The parameters of the simulation model were
obtained in the field and were calibrated using two
events: event A and B. The simulations performed
using the deterministic model show that the
simulated trajectories with DEM 1 resemble reality
much more than the simulated trajectories using
DEM 2. The trajectories of the two simulations (Fig.
2.11) show that the simulation of event A is much
more accurate using DEM 1 than DEM 2
(20mversus 90mof deviation of trajectory).
Likewise, in event B the accuracy usingDEM1 is
greater (trajectories being practically identical
versus maximum lateral deviation of 20 m). An
improved knowledge of trajectories and coordinates
of the detachment area enables us to better assess
the definition of areas exposed to rockfall.



2.5. DISCUSSION

The Terrestrial Laser Scanner (TLS) is an
instrument of great accuracy and speed in the
acquisition of 3D coordinates of millions of ground
surface points. Owing to the recent increase in its
Fig. 2.10. Contour lines (equidistance: 5 m.) plotted using DEM 1 (a) and DEM 2 (b): DEM 2 was obtained using digital
topographic map 1:5.000. DEM 1 was generated using point clouds acquired with TLS ilris-3D. Despite the fact that both DEM
have s similar resolution (6 m), DEM 1 is more accurate than DEM 2. Coordinates are local.







Abellán et al. (2006) Chapter 2: Applications of TLS to rockfall studies





Antonio Abellán Fernández Page 43

maximum range, the long-range TLS-Ilris3D can be
used in the characterization of large land surfaces
(ranging between 10 and 106 m2) and over
distances of several hundred meters (max. distance
in our study: 870 m.). Most studies on laser
scanning technology are focused on data acquisition
and processing (e.g. Beraldin et al., 2000; Staiger,
2003; Alba et al., 2005), rather that on the
development of new applications using point clouds.
Our results using TLS show how high quality (7
mm. at 50 m) and quantity point cloud (4·106 points)
help us to better estimate parameters in rockfall
studies.

Staiger (2003) shows how other measurement
systems such as the tacheometer, laser-radar, are
more accurate (0.5 mm and 0.05 mm, respectively)
than the Terrestrial Laser Scanner, but the
maximum distance of range and speed in the
acquisition of points are much lower. Furthermore,
TLS is capable of obtaining millions of 3D points in
few minutes, whereas the acquisition of the same
information using a total station would be
impractical. Other remote sensing methods of
greater range such as aerial LIDAR cover larger
areas than TLS but are less accurate: vertical
accuracy and data density are around 10 cm and 1 m,
respectively (McKean and Roering, 2004).
Moreover, with aerial LIDAR that offers a vertical
view, it is not possible to acquire high density
information in areas with very steep slopes. Some
works (e.g. Bitelli et al., 2004; Demir et al., 2004)
compared TLS with photogrammetric techniques
and they suggest that TLS could be a good
alternative to photogrammetry because it produces
faster, more accurate and less expensive DEMs.
However, both terrestrial measurements suffer from
the problems of shadow areas (as in our data
acquisition), especially in case of rockslope
horizontal surfaces. A combination of terrestrial and
aerial measures can resolve these problems (Ruiz et
al., 2004). Another difficulty with TLS is the data
processing of a huge amount of point clouds.
Further studies on the optimization of the data
density acquired in each scan are necessary. It
Fig. 2.11. Comparison of trajectories observed in the field (solid line) with simulations undertaken using DEM 1
(dashed line) and DEM 2 (dotted line) for events A and B. The highest accuracy is achieved with DEM 1.






SECTION I. Introduction Publication A: Eng. Geol., 88(3–4), 136–148



Page 44 PhD thesis

would be useful to answer the question: How many
points are needed to obtain accurate results.

A given number of points that belong to the same
joint allow us to calculate the orientation (dip
direction and dip angle) of the joints that
characterize the bedrock. Herda (1999) shows how
manual measurements with a compass can produce
a high strike standard deviation for shallow-dipping
joints. In order to reduce this uncertainty, Feng et al.
(2001) compares manual measurements (compass)
with total station measurements, and concludes that
the problem of apparent orientation measurements
can be considerably reduced with a total station.
Moreover, direct contact between operator and the
rock face is avoided. An automatic method for
discontinuity measurements of rock outcrops with
TLS has been investigated (Slob et al., 2002; Slob
and Hack, 2004). This methodology yields
satisfactory results when the facets of the rock
outcrop are determined by the discontinuities. In
some cases (as in our study area), the whole outcrop
is not defined by discontinuities with the result that
user-selection of the points of the same joint is
necessary. Thus in our work the joint orientations
obtained via user selection measurement were used
to calculate the volume of detached blocks.

Coordinates of the detachment point and volume of
detached blocks were estimated using traditional
methods (visual recognition) and also based on
point cloud approach. The results show how the
values of kinetic energy differed using the two
methodologies, the latter being more accurate than
the former.

As regards surface modelling, micro-topography
and 3D variations of the concavities and convexities
of the slope play an important role in the dynamics
of rockfall paths. This has been recently highlighted
in field test observations of released ophiolitic
blocks with volumes varying from 0.01 to 0.6 m3
(Giani et al., 2004). Also, Crosta and Agliardi
(2004) demonstrate how 3D variations of the slope
(higher values of roughness and lower values of cell
size of the DEM) increase lateral dispersion of
trajectories in rockfall simulations. In our study, we
show how high accuracy DEM can be employed to
make a detailed rockfall simulation and obtain more
accurate estimates of the trajectories. The results of
the rockfall simulation with a lower accuracy DEM
(DEM 2) did not correspond to the trajectories
observed in situ (Fig. 2.11), whereas the simulation
using a higher accuracy DEM (DEM 1, obtained
with TLS) yielded rockfall trajectories closer to
reality. The exact acquisition of the trajectories
seems unlikely. The observation of the reality
shows how the rockfall trajectories are always
different despite the fact that they originate from the
same detachment point and have the same volume.

Other studies (e.g. Hunter et al., 2003) show also
monitoring possibilities with Terrestrial Laser
Scanner. New data acquisition is being planned to
obtain the temporally spaced scans. The temporal
frequency, together with all the parameters obtained
in this work, will be crucial for a rockfall hazard
assessment of a steep rock slope at Vall de Núria.
Owing to the improvements made in the 3D data
acquisition and in the parameters discussed in this
paper, it appears that TLS will soon become a
widely used instrument in the characterization and
monitoring of landslides.



2.6. CONCLUSIONS

Long-range Terrestrial Laser Scanner is a tool that
allows the rapid, accurate and relatively simple
geometrical characterization of large areas of rock
slopes. This technology was successfully used at
Vall de Núria.

With the aid of the point cloud obtained with
TLSIlris3D it is possible to calculate an accurate
Digital Elevation Model and obtain a detailed
inventory of rockfalls. 3D data on the land surface
considerably improve geometrical characterization,
resulting in a better understanding of rockfall
phenomena. More accurate information on the
detachment location, joint orientation, volume of
detached blocks and characterization of rockfall
trajectories and energies are obtained. These
improvements enable us to better assess the areas
exposed to rockfalls and to provide more reliable
estimates of impact energy. These factors should be
borne in mind when hazard zoning is established
and when protective measures are proposed.



ACKNOWLEDGEMENTS

This study is supported by RISKNAT project
(SGR2001-00081) and by a pre-doctoral grant from
the Ministerio de Educación y Ciencia of Spain (ref.
no.: AP-2004-1852). Special thanks are due to
Altop- Metronic for help in data acquisition with
TLS-Ilris3D and to the Vall de Núria Company for
logistics support. The authors also wish to
acknowledge David Garcia, Dr. Felipe Buill and
Pau Arbues for their support in the data treatment.
We are indebted to George von Knorring and Giorgi




Abellán et al. (2006) Chapter 2: Applications of TLS to rockfall studies





Antonio Abellán Fernández Page 45

Khazaradze for improving the English version of
the manuscript. And finally, thanks are also due to
Drs. Janusz Wasowski, Vincenzo Del Gaudio,
Domenico Capolongo and an anonymous reviewer
for suggestions and critical inputs that helped to
improve the quality of the manuscript.



References

Alba, M., Longoni, L., Papini, M., Roncoroni, F., Scaioni, M.,
2005. Feasibility and problems of TLS in modelling rock
faces for hazard mapping. In: proceedings of the ISPRS
workshop laser scanning 2005, Enschede, Netherlands.
WG II/3, III/4, V/5. Available at
<www.commission3.isprs.org/laserscanning2005>.
Agliardi, F. Crosta, G.B., 2003. High resolution three-
dimensional numerical modelling of rockfalls.
International journal of Rock Mechanics & Mining
Sciences, Vol.40, 455-471.
Baillifard, F., Jaboyedoff, M. & Sartori, M., 2003. Rockfall
hazard mapping along a mountainous road in Switzerland
using an empirical approach. Natural Hazards and Earth
System Sciences. Vol. 3, 435-442.
Barton, N., Bandis, S., 1990. Review of predictive
capabilities of JRC-JCS model in engineering practise.
Proc.Int.Conference Rock Joints, Balkema publ.,
Rotterdam, 603-610.
Bates, R.L., Jackson, J.A., 1987. Glossary of Geology (third
edition). American Geological Institute 788p.
Beraldin, J.A., Blais, F., Boulanger, P., Cournoyer, L.,
Domey, J., El-Hakim, S.F., Godin, G., Rioux, M., Taylor,
J., 2000. Real World modelling through high resolution
digital 3D imaging of objects and structures. ISPRS
Journal of Photogrametry & Remote Sensing 55, 230-250.
Bieniawski, Z.T., 1993. Classification of rock masses for
engineering: the RMR system and future trends. In:
Comprehensive Rock Engineering, Hudson J.A. (ed.),
Pergamon Press, Vol. 3, 553-573.
Bitelli, G. Dubbini, M., Zanutta, A., 2004. “Terrestrial Laser
Scanning And Digital Photogrammetry Techniques To
Monitor Landslide Bodies”. In: proceedings of the XXth
ISPRS congress, Istanbul, Turkey. Commission V, WG
V/2.
Bronshtein, I.N., Semendyayev, Musiol, G., Muehlig, H.,
2003. Handbook of Mathematics, 4
th
ed.. Springer, Berlin,
1153 p.
Copons, R., Altimir, J. Amigó, J., Diaz, A., Vilaplana, J.M.,
2000. Eurobloc: un modelo de simulación de caída de
bloques y su máxima adaptación a la realidad. Geotemas,
Vol. 1, p.219-222.
Copons, R., Vilaplana, J.M., Corominas, J., Altimir, J.,
Amigo, J., 2005. Rockfall risk management in high
density urban areas. The andorran experience. In Glade,
T., et al. Landslide Hazard and Risk. John Wiley & Sons,
pp. 675-698.
Corominas J., Copons R., Vilaplana J.M., Altimir J., Amigó J.,
2003. Integrated Landslide Susceptibility Analysis and
Hazard Assessment in the Principality of Andorra.
Natural Hazards, November 2003, vol. 30, iss. 3, pp. 421-
435(15).
Crosta G. B. , F. Agliardi, 2004. Parametric evaluation of 3D
dispersion of rockfall trajectories. Natural Hazards and
Earth System Sciences, Vol. 4 (4) 583 – 598
Demir N., Bayram B., Alkis Z., Helvaci C., Çetin I., Vögtle
T., Ringle K., Steinle E., 2004. Laser Scanning for
Terrestrial Photogrammetry, Alternativesystem Or
Combined With Traditional System?. In: Proc of the
XXth ISPRS Congress, Istanbul, Turkey, commission V,
WG V/2. p. 193-198.
Dorren, L. 2003. A review of rockfall mechanics and
modelling approaches. Progress in Physical Geography
27,1 (2003) pp. 69–87
Dorren, L.K.A., Seijmonsbergen, A.C., 2004. Comparison of
three GIS-based models for predicting rockfall runout
zones at a regional scale. Geomorphology, Vol 56, Issues
1-2, pp.49-64
Feng, Q., Sjögren, P., Stephansson, O., Jing, L., 2001.
Measuring fracture orientation at exposed rock faces by
using a non-reflector total station. Engineering Geology
59, 133-146.
Fernández, M., Vilaplana, J. M., 2004. Factores
meteorológicos que desencadenan desprendimientos de
rocas en Vall de Núria. En: G. Benito & A. Díez Herrero
(Eds.), Riesgos Naturales y Antrópicos en Geomorfología,
Vol.II. Madrid: SEG – CSIC Centro de Ciencias
Medioambientales, 389 – 401
Fernández, O., 2005. Obtaining a best fitting plane through
3D georeferenced data. Journal of Structural Geology, 27,
855-858
Giani, G.P. Giacomini, A. Migliazza, M. and Segalini, A.,
2004. Experimental and Theoretical Studies to Improve
Rock Fall Analysis and Protection Work Design, Rock
Mech. Rock Eng. 37 (5), 369–389
Günter, A., 2003. Slopemap: programs for automated
mapping of geometrical and kinematical properties of
hard rock hill slopes, Computer & Geosciences 29/7, 865-
875.
Günther, A., Carstensen, A., Pohl, W., 2004. Automated
sliding susceptibility mapping of rock slopes. - Natural
Hazards and Earth System Sciences, Vol. 4, 95-102.
Guzzetti F., Reichenbach P. & Ghigi S., 2004. Rockfall
hazard and risk assessment in the Nera River Valley,
Umbria Region, central Italy. Environmental Management,
Vol. 34:2, 191-208
Hancock, J., Langer, D., Hebert, M., Sullivan, R., Ingimarson,
D., Hoffmann, E., Mettenleiter, M., Fröehlich, C., 1998.
Active laser radar for high-performance measurements.
Proceedings of the 1998 IEEE International Conference
on Robotics and Automation, Vol.2, 1465-1470.
Herda, H.H.W., 1999. Strike standard deviation for shallow-
dipping rock fracture sets. Rock Mechanics and Rock
Engineering, 32 (4), 241-255
Hoek, E., 2000. Analysis of rockfall hazards. Rock
Engineering: course notes, p. 115-136. Rocscience.
Available at <www.rocscience.com>.
Hunter, G., Pinkerton, H., Airey, R., Calvari, S., 2003. The
application of a long-range laser scanner for monitoring
volcanic activity on Mount Etna. Journal of Volcanology
and Geothermal Research 123, 203-210.
Marquínez J., Menéndez Duarte R., Farias P., Jimenez M.,
2003. Predictive GIS-Based Model of Rockfall Activity in
Mountain Cliffs. Natural Hazards, Vol. 30 (3), 341-360.
McKean, J., Roering, J., 2004. Objective landslide detection
and surface morphology mapping using high-resolution
airborne laser altimetry. Geomorphology, 57, pp.331–351.




SECTION I. Introduction Publication A: Eng. Geol., 88(3–4), 136–148



Page 46 PhD thesis

Palmström, A., 1995. RMi: A rock mass characterization
system for rock engineering purposes. PhD thesis, Oslo
University, Norway, 400p. Available at
<www.rockmass.net>.
Pfeiffer, T.J Higgins, J.D., Andrew, D.R., Schultz, R.J., Beck,
R.B., 1995. CRSP: Colorado Rockfall Simulation
Program, Version 3.0a User’s Manual. Colorado School
of Mines.
Piteau and Associated Limited, 1980. Slope stability analisys
for rockfall problems: The computer rockfall model for
simulating rockfall distribution. Rock Slope Engineering,
p. 62-68. Department of transportation, Washington DC.
Rendón, 2004. Evaluación y zonificación de la peligrosidad
geológica por desprendimientos en Vall de Núria,
Pirineos orientales España. PhD thesis (Annexe 1),
Universitat de Barcelona, Spain, 250p.
Rendón, A., Vilaplana, J.M, 2004. Propuesta para evaluar y
cartografiar la susceptibilidad a los desprendimientos de
ladera naturales. Geotemas 6(1), 359-362.
Rendón, A., Vilaplana, J.M., Copons, R., 2004. El análisis
volumétrico de los depósitos de talud: una etapa en la
evaluación de la peligrosidad por desprendimientos. En:
G.Benito & A. Díez Herrero (Eds.), Riesgos Naturales y
Antrópicos en Geomorfología, Vol.II. Madrid: SEG –
CSIC Centro de Ciencias Medioambientales, 413 – 421.
Romana, M. 1993. A Geomechanical Classification for
Slopes: Slope Mass Rating. In comprehensive rock
engineering, Hudson J.A. (ed.), Pergamon Press, Vol. 3.
Rouiller, J., Jaboyedoff, M., Marro, Ch., 1998. Pentes
instables dans le penique valaisan. Ed V/d/f. Zurich. 239p.
Rowlands, K.A., Jones, L.D. and Whitworth, M. 2003
Landslide laser scanning: A new look at an old problem.
Quarterly Journal of Engineering Geology and Hydrology
36 (2), 155 - 157.
Ruiz, A. Kornus, W. Talaya, J., Colomer, J.L., 2004. “Terrain
Modeling in an Extremely Steep Mountain: a
Combination of Airborne and Terrestrial Lidar”, In: proc.
of the XXth ISPRS congress, Istanbul, Turkey. Vol.
XXXV, part B3 Com. III,
Santanach, P., 1974. Estudi tectònic del paleozoic inferior del
Pirineu entre la Cerdanya i el riu Ter. Ed. Fundació
Salvador Vives Casajuana, Barcelona, 133p.
Scioldo, G. (1991): Rotomap: analisi statistica del
rotolamento dei massi. Guida informatica ambientale.
Patron, Milano, 81–84.
Sequeira, V., Fiocco, M., Bostrom, G., Gonçalves, J.G.M.,
2003. 3D Verification of Plant Design. 25th ESARDA
Symposium on Safeguards and Nuclear Material
Management, Stockholm, Sweden. Available at:
<http://sir.jrc.it/sir/Publications/040130-DIV-
paper%20ESARDA.pdf>
Singh, S.West, J., 1991. “Cyclone: A Laser Scanner for
Mobile Robot Navigation”, Carnegie Mellon University,
Robotics Institute Technical Report, CMU-RI-TR-91-18.
Slob, S., Hack, H.R.G.K. and Turner, K., 2002. An approach
to automate discontinuity measurements of rock faces
using laser scanning techniques. In: Proceedings of ISRM
EUROCK 2002. Ed. by C. Dinid da Gama and L. Riberia
e Sousa. Lisboa, Sociedade Portuguesa de Geotecnia, pp.
87-94.
Slob, S., Hack, R., 2004. 3D Terrestrial Laser Scanning as a
new Field Measurement and Monitoring Technique. In:
Hack, R., Azzam, R., Charlier, R. (eds), Engineering
Geology for Infrastructure Planning in Europe: A
European perspective. Springer, Berlin. Lecture notes in
Earth Sciences, n.104, pp179-189.
Staiger, R., 2003. Terrestrial laser scanning technology,
systems and applications. 2nd FIG regional Conference,
Morocco. Available at <http://www.fig.net/pub/
morocco/proceedings/TS12/TS12_3_ staiger.pdf>
Varnes, D.J., 1984, Landslide Hazard Zonation: a Review of
Principles and Practice. Commission on Landslides of
IAEG, UNESCO, Paris, Natural Hazards, 3, 61 p.




Chapter 3: Summary of the methodology


Antonio Abellán Fernández Page 47




II
SECTION
Change
detection



SECTION II. Change detection



Page 48 PhD Thesis




Chapter 3: Summary of the methodology


Antonio Abellán Fernández Page 49
























CHAPTER 3:
Material and Methods







SECTION II. Change detection



Page 50 PhD Thesis





























Chapter 3: Summary of the methodology


Antonio Abellán Fernández Page 51















3.1 Material

In an attempt to contribute to the overall comprehension of the PhD thesis and to avoid unnecessary
repetitions of the text, this section refers to the material and methods used in subsequent chapters.


3.1.1 Instruments: TLS ILRIS-3D
The remote sensing tool employed in this study was a Terrestrial Laser Scanner (TLS). This
instrument is also known as a ground-based LIDAR (Light Detection and Ranging). We used an
ILRIS3D model (Optech). This instrument consists of a transmitter/receiver of infrared laser pulses
(1535 nm wavelength) and a scanning device. A brief description of the instrument was given in
Publication A (Section 2.1.2) and Publication B (Section 4.2.1). The number of publications on the
use of a TLS in earth science research has increased considerably in the last 5 years (see Section
8.1.2). Basic principles of laser ranging, profiling and scanning are discussed by different authors
(e.g. Lichti D.D., Jamtsho, 2006; Teza et al., 2007; Petrie and Toth, 2008). A study of the physics of
the laser beam, energy source, etc. is beyond the scope of this PhD thesis.




SECTION II. Change detection



Page 52 PhD Thesis
3.1.2 Data acquisition
The following two basic figures (not available in the compendium of publications) contribute to an
understanding of the data acquisition process: (a) Fig. 3.1 shows the basic operation of a laser
rangefinder based on the measurement of the time of flight (TOF) method; (b) Figure 3.2 shows the
shape of the waveform of the returned pulse. As can be seen in this figure, range measurements can
be undertaken using the first or last pulse of the return signal. We used the last pulse to obtain the
return signal from the rock face (in place of vegetation).





Fig. 3.1 Basic operation of a laser rangefinder (from Petrie and Toth, 2008)
























Fig. 3.2 Shape of the complete waveform of the returned pulse. The first and last pulse
can be seen in this picture (from Petrie and Toth, 2008; original figure by Brenner, 2006)




Chapter 3: Summary of the methodology


Antonio Abellán Fernández Page 53



Fig. 3.3 Data acquisition using a TLS ILRIS-3D (Optech)


3.2 Method

3.2.1 Comparison of sequential TLS datasets
Multi-temporal TLS data acquisition allows the geomorphological evolution of a given area to be
studied, such as rockfalls and small displacements (sections next section). The detection of minor
scale rockfalls is based on the method described in Rosser et al. (2005) and Lim et al., (2006). A
description of the alignment method and the calculation of the differences in each point of the TLS
datasets are provided in Publication B (Section 4.2.3). The influence of the density of points on the
quality of the geometric modelling is discussed in the same publication (Section 4.2.4). This section
also refers to an assessment of the quality of the model comparison using optimal point spacing
(OPS).


3.2.2 Detection of small displacements
Errors in sequential comparison of TLS datasets can potentially be reduced by using the information
of the neighbouring points, i.e. by filtering or interpolation averaging (e.g. Lindenbergh and Pfeifer,
2005; Araiba, 2006). The measurement of precursory displacement was considerably improved by
applying nearest neighbour averaging (NN), which reduces the error by up to a factor of 6 (see
Chapter 5). The experiment and the displacement computation are described in Sections 5.2.2 and
5.2.3 respectively. A comparison of this technique with a conventional RAW comparison is
discussed in the same publication (Sections 5.3.1 and 5.3.2).


3.2.3 Prediction of rockfalls
Rockfall predictions are usually based on the measurement of some indicators before failure, e.g.
crack opening, acoustic disturbances, microseismicity, and/or precursory deformation. The



SECTION II. Change detection



Page 54 PhD Thesis
prediction of rockfalls in this research is based on the detection of the latter (slope creep or
precursory deformation, Terzaghi, 1950). The aforementioned method (Section 3.2.2) enabled more
accurate detection of this deformation. Publication C describes the application of this method to a
50m
3
rockfall through a back analysis procedure (Section 5.4). Further methodological issues are
discussed in Publication D (Section 6.2.4). The early detection of precursory deformation in four
additional case studies is analyzed in Chapter 7. New datasets acquired during 2009 confirmed the
evolution of precursory deformation in these ongoing displacements and revealed the existence of
new areas (see Chapter 8).

























































CHAPTER 4
Rockfall monitoring at
Castellfollit de la Roca

Publication B: Abellán et al., (submitted). Nat. Hazards Earth Syst. Sci





























































Nat. Hazards Earth Syst. Sci.
www.nathazardsearthsystsci.net
© Author(s). This work is distributed under
the Creative Commons Attribution 3.0 License.











Rockfall monitoring by Terrestrial Laser Scanning.
Case study of the basaltic rock face at Castellfollit de
la Roca (Catalonia, Spain).
Abellán, A.
1
, Vilaplana, J. M.
1
, Calvet, J.
1
, García-Sellés, D.
1
, Asensio, E.
1

1
RISKNAT group & GEOMODELS Institute, Depart. of Geodynamics and Geophysics, Univ. of Barcelona, Spain

Received: 5 July 2009

Abstract. This case study deals with a quaternary basaltic rock face monitoring using a relatively
new remote sensing technique: a Terrestrial Laser Scanner (TLS). The pilot study area is an almost
vertical, fifty meters high cliff on top of which the village of Castellfollit de la Roca is located.
Rockfall activity is currently causing a retreat of the rock face, which may endanger the houses
located at its edge. TLS datasets consist of high density 3D point clouds acquired 9 times in a time
span of 22 months (from March 2006 to January 2008). The change detection, i.e. rockfalls, was
performed through a sequential comparison of datasets. The quality of the rock face modelling was
assessed for different densities of points, obtaining the optimum point spacing between 4.5 and 5.5 cm.
Two types of mass movement were detected in the monitoring period: (a) detachment of single basaltic
columns, with magnitudes below 1.5 m3 and (b) detachment of group of columns, with magnitudes
from 1.5 to 150 m3. Furthermore, the historical record revealed the occurrence of slab failures with
magnitudes higher than 150 m3. Displacements of a likely slab failure were measured, suggesting an
apparent stationary stage. Our results together with the study of the historical record enabled us to
estimate the mean rate of retreat of the cliff: from 6 to 11 mm/year. The application of the TLS
considerably improved our understanding of rockfall phenomena in the study area.



Keywords: Laser Scanning, monitoring, rockfall detection, toppling, columnar basalt
























Abellán et al. (submitted) Chapter 4: Rockfall monitoring at Castellfollit de la Roca



Antonio Abellán Fernández Page 59

4.1 INTRODUCTION

A rockfall, which is a fragment of rock detached by
sliding, toppling or falling, falls along a vertical or
sub-vertical cliff and proceeds down slope by
bouncing, flying or rolling (Varnes, 1978). Minor
scale rockfalls (up to several hundred of cubic
meters) are the most frequent type of landslide on
steep slopes in mountain areas (Copons and
Vilaplana, 2008), marine cliffs (Rosser et al., 2005)
and rock faces. Magnitude, frequency and velocity
of the rockfalls are the main parameters for a
rockfall hazard assessment (Varnes, 1978).

The possibility of acquiring datasets of the terrain
surface with a high accuracy and high spatial
resolution, either using laser, optical and/or radar
sensors mounted on terrestrial, aerial and/or satellite
equipment are currently opening up new ways to
visualize, model and interpret Earth surface
processes. Ground based sensor, e.g. Terrestrial
Laser Scanner (TLS), obtain its maximum resolution
on vertical surfaces, such as mountainous rock faces,
marine cliffs, etc.

TLS is one of the most promising remote sensing
techniques for rockslope characterization and
monitoring (Biasion et al., 2005; Bauer et al., 2005;
Oppikofer et al., 2008) because of its capability to
accurately acquire dense 3D coordinates of the
terrain. TLS equipment is currently being used by
different groups in the monitoring and hazard
assessment of slope movements. Its main
applications concerns the characterization of 3D
discontinuities (Jaboyedoff et al., 2007;
Sturzenegger and Otead, 2009) and change
detection, e.g. rockfalls (Lim et al., 2006; Rosser et
al, 2007), rock avalanches (Dunning et al., 2009)
and soil slides (Teza et al., 2007; Monserrat and
Crosetto, 2008). Recent studies highlight the
applicability of TLS to the estimation of the rates of
retreat of different rock slopes: Rabatel et al. (2008)
quantified the volume of the main rockfalls on the
east face of the Tour Ronde, Mont Blanc massif;
Oppikofer et al. (2008) discussed the movement of
various blocks and the final collapse of a rock spur
on the eastern flank of the Eiger peak (Swiss Alps);
Lim et al., (2009) studied the erosion rates of a
rocky coastal cliff at Staithes, North Yorkshire (UK)
and the influence of environmental variables during
a two year monitoring period. Finally, Dewez et al.
(2009) analyzed the erosion rates on a 700m long
coastal chalk cliff in Normandy.

In our study, this relatively new technology was
applied to the quaternary basaltic rock face at
Castellfollit de la Roca (Catalonia, Spain, Fig. 4.1a).
From a geomorphological point of view, rockfalls
currently constitute the main process at the rock face.
Fig. 4.1. (a) Location of the village and the basaltic formation at Castellfollit de la Roca (Garrotxa Volcanic Field, NE Spain);
(b) Perspective view with indication of the scanned rockface; (c) Synthetic stratigraphic sequence modified from Palli and Trilla
(1976). Thickness ~45m. See Table 1 for a description of the geological levels (from A to I); (d) Panoramic view of the North
face of the basaltic formation at Castellfollit de la Roca. Houses located on the edge of the cliff are visible.





SECTION II : Change detection Publication B: Nat. Hazards Earth Syst. Sci.


Page 60 PhD thesis

The rock fall and hence the rock face retreat could
pose an important risk to the houses located on the
edge of the cliff. The aims of the study are: (a) to
detect and characterize rockfalls during the
monitoring period through a comparison of
sequential TLS datasets; (b) to establish a cliff
retreat rate. Moreover, a metric scale crack was
detected parallel to the rock face. This crack
constitutes the detachment area of a probable rock
slab. As a consequence, the following aim was
added to the previous list: (c) to determine the rate
of crack opening during the TLS monitoring period.


4.1.1 Study area
The village of Castellfollit de la Roca is located at
the top of a basaltic formation bounded by two
scarps (see Fig 4.1a). The Rock face is currently one
of the main geomorphological highlights of the
Natural Park of the Garrotxa Volcanic Field (GVF).
The basaltic elevation is located between two rivers:
river Fluvià towards the North and river Turonell
towards the south. The basaltic formation is made up
of two lava flows (table 4.1 and Fig. 4.1c): Upper
lava flow (levels A, B, C, D, table 4.1) and Lower
lava flow (levels E, F, G). Dating of these flows by
K-Ar method yielded ages of 192.000 ± 25.000 and
LEVEL THICKNESS
GEOLOGICAL
DESCRIPTION
(1)

GEOMECHANICAL
DESCRIPTION
(2)

(A) 8~10 m Basalts. Massive structure
Partially weathered and transformed into a
soil in the upper part (Ros et al., 1996). A
weakening of the geomechanical properties
and an increase in porosity is to be
expected.
(B) 8~10 m
Basalts. Columnar jointing:
hexagonal pattern.

(C) 5 m Basalts. Wavy prismatic structure
(D)
Upper

lava

flow
5 m
Basalts. Columnar jointing:
hexagonal pattern (same as B
level).

Discontinuities result from contraction upon
cooling of the lava flow, giving rise to single
columns. These discontinuities control the
stability of the blocks. Overhanging parts
may affect the local stability of the rock face.
(P)
Paleo-
soil
0.5~1.5 m
Irregular layer of clays and
pyroclasts in the lower part. The
upper part contains a paleosoil
and unconsolidated sediments.

Weakness level (low mechanical properties).
Pyroclasts are not cemented together.
Furthermore, the paleosoil in the upper part
is unconsolidated. In this layer, porosity is
higher than in other layers.
(E) 3~4 m
Basalts. Wavy prismatic structure.
In some areas also with a radial
structure.
Same as B, C and D levels
(F) 4~5 m Basalts. Lenticular structure.
High weathering. Discontinuities control the
stability of the blocks. Overhanging parts
may affect the local stability of the rock face.
(G)
Lower

lava

flow
1.5 m
Basalts. Prismatic layer of short
columns.

(H) Fluvial -
Quaternary fluvial deposits prior to
lava flows.

(I) Bedrock -
Bedrock formed by Eocene
sandstone

Weakness level (low mechanical properties).
The erosion of the river may affect global
stability because of undermining
phenomena.
(1)
Composition, texture, structure, etc.
(2)
Mechanical behaviour, porosity, weathering, etc.
Vegetation is scattered at all the levels and some cracks are clay filled. Crack opening due to vegetation could be a
conditioning factor. Rainfall and pos/neg. variations of temperature are relatively frequent in the study area. As a
result, crack opening by gelifraction seems to be a relatively common process in the study area.

Table 4.1. Main geological and geomechanical characteristics of the stratigraphic sequence.





Abellán et al. (submitted) Chapter 4: Rockfall monitoring at Castellfollit de la Roca



Antonio Abellán Fernández Page 61

217.000 ± 35.000 years BP, respectively (Donville,
1973; ICC et al., 2007). These units are linked by an
irregular deposit formed by pyroclasts and a
paleosoil (Pallí and Trilla, 1976; Mallarach and
Riera 1981). Table 4.1 gives a description of the
different layers from the viewpoints of geology and
engineering geology (modified form Pallí and Trilla,
1976; Mallarach and Riera, 1981; Martí et al., 2000;
Mascort et al., 2004; ICC et al., 2007).


4.1.2 Historical inventory of rockfalls
Mass movements in the study area are classified as
follows: (a) detachment of single columns (Fig.
4.2a). The magnitude of this type of event is
characterized by volumes below 1.5 m
3
. As
discussed below, our results show a frequency
higher than that in the historical record; (b)
detachment of a group of columns with a magnitude
from 1.5 to 150 m
3
(Fig. 4.2b). This type of failure
may involve dozens of columns and may affect
different levels. The historical record shows 2
events in the last 30 years; (c) Rock slab failures
with magnitudes higher than 150 m
3
(Fig. 4.2c).
Two slab failures occurred in the last 30 years.

Table 4.2 shows the historical record in the 30 years
prior to the monitoring period: 1976-2006.

Rockfall phenomena in the study area are poorly
documented: there are few data on back analysis of
prior events (Ros et al., 2006; Pallí and Trilla, 1976;
Mallarach and Mirabell, 1976; Ros et al., 1996;
Culebras, 2002), magnitude-frequency relationships,
conditioning and triggering factors, susceptibility
assessment (Asensio et al., in press). Population
surveys were conducted to confirm the low
frequency of great magnitude rockfalls in the last 50
years. The combination of the historical record
(table 4.2) and the population surveys is referred to
as a long-term approach below.


4.1.3 Conditioning and triggering factors
The role of conditioning factors such as jointing
pattern, lithology and morphology of the cliff is
discussed below.

(a) The pre-existing columnar jointing pattern
played a key role in the geometry of the detached
blocks. Evidence for this was provided by the
rockfalls detected regardless of their size. The
influence of the meso-scale structure, i.e. columnar
pattern, in the macro-scale morphology of the rock
face was also observed: the mean orientation of the
rock face corresponds to the mean orientation of the
facets of the basaltic columns; (b) The different
layout and composition of each of the levels
described in Table 4.1 also played an important part
in the evolution of the rock face. Two of the levels
described in Table 4.1 were identified as weak levels:
pyroclastic level and fluvial deposits. The erosion
and/or weathering of these levels create an overhang
in the upper levels, affecting their stability, as in the
Fig. 4.2. Type of rockfall according to its volume: (a) detachment of single columns (volume below 1.5 m
3
); (b) detachment of a
group of columns (volume from 1.5 to 150 m
3
); (c) likely slab failure (volume higher than 150 m
3
).





SECTION II : Change detection Publication B: Nat. Hazards Earth Syst. Sci.


Page 62 PhD thesis

case of the interconnected rockfalls that occurred in
February and September of 1976 (table 4.2). The
former rockfall was preceded by a great magnitude
event in the lower part of the slope; (c) The
morphology of the cliff may affect the local stability
of the rock face, e.g. the protruding parts of the
slope are likely to collapse.

The role of the triggering factors, i.e. climatic
variables, undercutting and earthquakes is described
as follows:

(a) As regards climatic variables, earlier works (Ros
et al., 1996; Pallí and Trilla, 1976) in addition to
population surveys have drawn attention to the role
of precipitation in the rockfalls in the last 30 years
(table 4.2). Daily rainfall and temperatures are
recorded by the Olot meteorological station (Servei
Meteorològic de Catalunya, www.meteo.cat),
located 7.8 km to the SW of the study area. This
record shows a relatively high rainfall occurrence
(1000 mm/year) and a mean of 50 positive/negative
variations of the temperature in a given year.
Although no direct relationships between climatic
variables and rockfall occurrence are discussed in
our research, a crack opening due to an ice wedge
could be a relatively common process.
(b) Another triggering factor may be the erosion of
the scree deposits and/or of the lower part of the
rock face in extraordinary floods. This hypothesis
was supported by Pallí and Trilla (1976) and Ros et
al. (1996). The maximum curvature of the rock face
corresponds to the meander, i.e. the concave part of
the river (Fig.4.1), where there is more erosion. This
hypothesis may be supported by the fact that the two
main mass movements of the historical record (table
4.2) took place in this area; (c) Finally, the study
area is located in a moderate seismic area. Despite
its low recurrence, seismicity could also be regarded
as a triggering factor, e.g., a destructive earthquake


DATE OF
THE
ROCKFALL
LOCATION
(SOUCE AREA)
VOLUME
(1)

DAMAGE TRIGGERING REFERENCES
1976, Feb.
Area of concavity of
the North face, lower
part.
960m
3

Any damage.
Undercutting of
the toe of the
slope.
Continuous
rainfall
Palau (1976);
Mallarach and
Mirabell (1976)
1976, Sept.
Area of concavity of
the North face,
central and upper
part of the rock face.
1500m
3

Structural
damage to the
porch of a
house
Area destabilized
by the previous
rockfall
Palli and Trilla
(1976); Mallarach
and Mirabell
(1976)
1977 Unknown. 5 m
3

No severe
damage
unknown
Mallarach and
Mirabell (1976)
1995, Nov.
Area of concavity of
the North face, upper
part.
50 m
3

Structural
damage to the
porch of a
house
Heavy rainfall Ros et al. (1996)
2001, March
Single column failure
in the central part of
the North face.
1 m
3
No damage unknown Culebras (2002)
2005, Feb.
SE part of the rock
face.
1 m
3
No damage Heavy rainfall
2006, Gen/Feb.
2 single columns in
the SE part of the
rock face
2 to 3 m
3
(1 to 1.5 m
3
each)
No damage Heavy rainfall

Population survey
conducted by
RISKNAT research
group

(1)
Volumes indicated in this table were roughly estimated from scar dimensions by the aforementioned authors.
Table 4.2. Historical record of rockfalls in the 30 years prior to our study.






Abellán et al. (submitted) Chapter 4: Rockfall monitoring at Castellfollit de la Roca



Antonio Abellán Fernández Page 63

causing 85 fatalities at the village of Castellfollit de
la Roca occurred on 2 February 1428 (Olivera et al.,
2006).



4. 2 MATERIALS AND METHODS

4.2.1 Terrestrial Laser Scanner (TLS)
The remote sensing tool employed in this study is a
TLS. This instrument is also known as a Ground
based LIDAR (Light Detection and Ranging). We
used an ILRIS3D model (Intelligent Laser Ranging
and Imaging System), from Optech. The instrument
consists of a transmitter/receiver of infrared laser
pulses (1535 nm wavelength) and a scanning device.
The laser beam is directly reflected on the land
surface, obviating the need for intermediate prism
reflectors. TLS ILRIS3D shows a very high data
acquisition speed (up to 2,500 points/second)
compared with conventional surveying methods, e.g.
total stations. Distance measurement (range, ρ) is
based on the Time-Of-Flight (TOF) of the laser
pulse (equation 1):

ρ = c·(TOF/2)

[equation 1, Petrie and Toth, 2008]
Where c = speed of light

Range measurement can be undertaken using first or
last pulse of the return signal. This range is
measured for an area equal to the spot dimension
(SD). This parameter is affected by a laser beam
divergence process (equation 2):

SD= ρ·tanα + a
[equation 2, Petrie and Toth, 2008]

Where α = angle of divergence (0.00974º); a=initial
beam size (12 mm); Equation in any consistent units
of SD and ρ.

The origin of the Cartesian Coordinate System (P0
=0,0,0) is set at the centre of the TLS instrument.
Coordinates of each point are acquired in a polar
system (ρ: range; ϑ: horizontal angle; ϕ: vertical
angle). The binary file provided by TLS was
transformed into a Cartesian system (x, y, z) using
Parser v 4.3.5.4. As a result, a Parametric Image
Format (PIF) file was obtained.

The scanning device uses an internal system of
rotating mirrors. Spatial resolution is obtained as a
function of the range and the angular resolution. The
latter is a function of the point spacing and the spot
dimension (Lichti and Jamtsho, 2006). In addition to
the 3D coordinates, the device is able to acquire the
value of intensity (l) of reflectivity of the land
surface on which the laser beam is focused. This
parameter is defined as the amount of reflected
signal with respect to the emitted one. It mainly
depends on the range, angle of incidence, moisture
and object material. Its value is usually normalized
on a 0-255 scale.

Although technical characteristics supplied by the
manufacturer show a high point accuracy (σ~0.7 cm
at 100m), this value is significantly worse in “real
conditions”. The accuracy of the measurement is a
main function of the range, reflectivity of the
material (Voegtle et al., 2008), complexity of the
scanned surface (Abellán et al., 2009) and angle of
incidence (Lichti, 2007). We tested our device at
different geological sites and obtained a maximum
optimal range of around 600 meters for dry surfaces
(e.g. Abellán et al., 2006; Vilajosana et al., 2008;
Abellán et al., accepted). Polyworks© v9.0 was the
main software used for the visualization and
treatment of the point clouds. Alignment and
comparison of the datasets was performed by the
ImAlign & Inspect modules.


4.2.2 Data acquisition
The first dataset (referred to as reference point cloud,
R0 below) was acquired in March 2006. The
datasets were acquired from five stations (portion of
the scanned face, perspective): (a) Station A (North
face, frontal); (b) Station B (North-East face,
oblique); (c) Station C (East face, oblique); (d)
Station D (South-East face, frontal) and (e) Station E
(North-West, oblique). From these sites, occluded
areas were minimized, allowing for the alignment
and merging in a single file, creating a final 3D
model (Fig. 4.3). A minimum of seven point clouds
was necessary to cover the whole rock face. As
discussed below, the mean point spacing of the
acquired datasets ranged from 4.5 to 5.5 cm. Each
point of the R0 was defined as a node for a Triangle
Irregular Network (TIN) surface (*.pif file). This
surface will be referred to as the surface of reference
(S0) below. The last pulse of the laser signal was
acquired in order to obtain the return signal of the
rock face (in place of vegetation). Data acquisition
was repeated 94, 186, 260, 368, 382, 443, 561, 662
days after R0 (periods i, ii, iii, iv, v, vi, vii and viii,
respectively). Each of these TLS datasets is referred
to as data point cloud (D1, D2, D3 … Dn) below.





SECTION II : Change detection Publication B: Nat. Hazards Earth Syst. Sci.


Page 64 PhD thesis

4.2.3 Comparison of sequential TLS
datasets
3D temporal variations of the terrain were detected
by comparing sequential datasets in accordance with
the methodology described in Rosser et al. (2005)
and Lim et al. (2006). A brief description is as
follows: (a) acquisition of R0 (see previous section);
(b) construction of the S0; (c) acquisition of D1,
D2,….. Dn; (d) alignment of these datasets with S0
(explained below); (e) comparison between S0 and
successive Di and (f) calculation of the differences
for each period of comparison.

The roto-translation parameters of the alignment
matrix (step d) were obtained in three stages: (a) a
preliminary registration was performed by a visual
identification of homologous points; (b) the
alignment was subsequently optimized using an
Iterative Closest Points (ICP) procedure (Chen and
Medioni, 1992). Using this algorithm, the
differences between points were progressively
reduced by a minimization of a mean square cost
function; and (c) the final improvement was
obtained by progressively reducing the “search
distance” parameter up to a few centimetres. As a
consequence, the metric scale changes (e.g. rockfall)
that occurred in the different intervals did not affect
the global alignment, alignment error being
negligible. Apart from the recorded metric scale
rockfalls, the surface of the rock face remained
practically unchangeable during the scan
comparison. The exact date of each event is an
unknown factor because of the non continuous data
acquisition and the absence of eyewitnesses.

The single point distances between the S0 and the
subsequent Di were computed in the ImInspect
module of PolyWorks© v.9 software using a
conventional methodology (data vs. reference
comparison). Comparisons based on “Shortest
distance” methodologies reduced the values of the
real changes that took place, i.e. rockfalls. Hence,
the direction of comparison was defined as the
normal vector of the rock face at its central part (Y
axis). Differences (Difi) were calculated for each
point as shown in equation 3:

Difi = Distance [SiP0]- Distance[S0P0]
[equation 3]

Part of the value of Difi is due to systematic
(instrumental and methodological) errors. The other
part is due to “real changes” in certain parts of the
slope. As regards the sign criteria, we use positive
values when the time of flight of the laser signal for
Di is higher than that of S0. As a result, positive
values correspond to a lack of material at a given
point, i.e. detachment of the material. Likewise,
negative values correspond to an increase in material,
i.e. scree deposits, or a displacement towards the
origin of coordinates. This negative displacement
may also reflect the pre-failure deformation in a part
of the slope (i.e. 2009; Abellán et al., accepted). The
calculation of the volume of the main events was
made using the “surface to plane” command of
Fig. 4.3. Images of the aligned TLS datasets: (a) left view: “-X axis”; (b) zenithal view: “+Z axis”; (c) frontal view: “–Y axis”.







Abellán et al. (submitted) Chapter 4: Rockfall monitoring at Castellfollit de la Roca



Antonio Abellán Fernández Page 65

IMInspect module (PolyWorks). In line with Rosser
et al., (2005), rockfalls with a volume under 0.001
m
3
were not considered in this study.


4.2.4 Assessment of the quality of the
rock face modelling
The influence of the angular resolution, i.e. density
of points, in the quality of the surface of reference
was tested by comparing two point clouds acquired
consecutively. Fig. 4.4a shows the percentiles of the
error as a function of the density of points. As
expected, the smaller the density of points, the lower
the accuracy of the comparison. The value of the
75th, 90th and 95th percentiles show similar values
of the error for a point spacing below 4.7 cm. As a
result, no significant improvements are obtained by
increasing the density of points. For this reason, the
Optimal Point Spacing (OPS) ranges from 4.5 up to
5.5cm in the study area. This value was selected for
the subsequent data acquisition campaigns.

Figure 4.4b shows the histogram of the model
comparison in the same section, using a point
spacing of 4.7cm. This histogram is characterized by
a very high kurtosis, i.e. the presence of infrequent
extreme deviations. Assuming a normal distribution,
the dispersion of the 68% of the population can be
explained by the standard deviation (1σ) parameter.
However, in accordance with ASPRS LiDAR
Committee (2004), a normal distribution of error
cannot be assumed, as in the case of Fig. 4.3b, where
the population does not fit (by far) a
normal/Gaussian distribution. Alternatively, and as
stated above, the variance of the population was
assessed by means the percentiles of the error. Two
different populations can be found in this histogram:
(a) Instrumental error, corresponding to the error of
90% of the population, i.e. between -3.04 and +2.99
cm (5th and 95th percentile, respectively); (b)
Outliers: data artefacts that were mainly found along
the boundaries of the occluded parts of the slope
with respect to the line-of-sight (LOS). These data
artefacts were quantified by the 99th percentile, i.e.
the error of 2% of the population was higher than
8.39 cm. The morphology of the rock face plays an
important role in the magnitude and extent of these
outliers: the greater the complexity of the scanned
surface, the higher the variance of the measurement
(Abellán et al., 2009). In the study area the outliers
tend to concentrate along the borders of the basaltic
columns.

The quality of the measurement can be improved by
filtering the areas similarly oriented to the LOS and
by the deletion of erroneous points, i.e. vegetation,
wires, birds, unexpected points, etc. The most
accurate results were obtained from stations A and D,
respectively. Jaszczack (2006) reported the accuracy
of the TLS point cloud for the different geological
layers and sections of the rock face.



Fig. 4.4. Assessment of the quality of the rock face modelling. (a) Variance of the population assessed for different densities of
points through the percentiles of the error of the model comparison; (b) Probability distribution of the error of the model
comparison using a mean point spacing of 4.7cm. The histogram is characterized by considerable kurtosis and a poor
resemblance to a fitted Gaussian distribution (blue line). See text for a detailed explanation.





SECTION II : Change detection Publication B: Nat. Hazards Earth Syst. Sci.


Page 66 PhD thesis

4. 3 RESULTS

4.3.1 Detachment of single columns
The detachment of single columns is the mass
movement of highest frequency and lowest
magnitude in the study area. The detachment of 6
basaltic columns was recorded in the 22 month
monitoring period. The magnitude of each of these
rockfalls is below 1.5 m
3
. The geometry of these
rockfalls is controlled by the columnar jointing
pattern. Fig. 4.5 shows a sequential comparison of
September and December 2006 TLS datasets in the
southern part of the rock face. A single column
detachment is clearly visible in the middle of the
figure.


4.3.2 Detachment of a group of columns
A failure that affected a group of columns in the
central part of the rock face (Fig. 4.6a) was recorded
in our research. This event can be described as a
combination of two interconnected detachments: (a)
the April 2007 rockfall (event i, Fig. 4.6b) is the
event of the highest magnitude in the monitoring
period. This figure shows a comparison of the
sequential datasets corresponding to 30 May 2007
and 13 April 2007. An irregular failure with
maximum dimensions of 15m height, 6m width and
1.5 thickness is observed in the middle of the figure.
Tens of basaltic columns belonging to levels B, C
and D were mobilized. The rockfall volume (50 m
3
)
was calculated by volume differences between pre
and post failure surfaces. The morphology of the
cliff prior to the rockfall revealed a protruding block
partly supported by the pyroclastic (P) level.
The geometry of the surface of detachment is
controlled by the columnar jointing pattern. A period
of continuous rainfall (100mm in 6 days) may have
triggered this rockfall; (b) a second event (event ii,
Fig. 4.6c) took place in this area six to nine months
after the rockfall of April 2007. This event consisted
of a few blocks from level C (1.5 m
3
) in addition to a
single column from level B (1.5 m
3
). As in the case
of the interconnected rockfalls of February and
September 1976, the origin of this event was
interpreted as a gravitational readjustment of the
scarp after the event of April 2007. As a result of
these rockfalls, the upper part of the slope lost part
of its basal support. Hence, the occurrence of a
future event in this area is likely.


4.3.3 Slab failures
No rock slab failures were recorded in the
monitoring period. However, the detection of a
metric scale crack during fieldwork (Fig 4.2.c)
provided evidence of the occurrence of an ongoing
rock slab failure with an estimated volume
exceeding 500 m
3
. This section of the rock face was



Fig. 4.5. Comparison of the sequential TLS datasets. Each point of the 3D point cloud is colour-coded in accordance with the
changes recorded during the period of comparison (September vs. December 2006). The failure of a single basaltic column and
small changes in vegetation are visible. Pre and post failure stages are visible on the left.






Abellán et al. (submitted) Chapter 4: Rockfall monitoring at Castellfollit de la Roca



Antonio Abellán Fernández Page 67


Fig. 4.6. (a) TIN surface of a part of the rock face where events i and ii took place; (b) Comparison of the March and April
2007 datasets; (c) Comparison of the October 2007 and January 2008 datasets. Each point is colour-coded according to the
changes during the period of comparison. Colour scale indicates positive differences along the Y direction, e.g. rockfalls.
Note that event ii is in the same area as event i.





SECTION II : Change detection Publication B: Nat. Hazards Earth Syst. Sci.


Page 68 PhD thesis

monitored by TLS given the high vulnerability of
the houses located on the edge of the cliff. The
results showed no significant pre-failure
deformation in this area in the 22 month monitoring
period. This suggests that the maximum value of the
displacement is lower than the 95th percentile of the
instrumental error by the time span of the
monitoring period, i.e. lower than 1.64 cm/year (3
cm in 22 months). Although this hypothetical
displacement seems currently stationary, its future
reactivation cannot be ruled out.



4. 4. DISCUSSION

4.4.1 Pros and cons of TLS
The use of TLS enabled us to better understand
rockfall phenomena of the rock face at Castellfollit
de la Roca. While classic surveying methods (i.e. a
total station, extensometers, etc.) allow for the
acquisition of millimetric accuracy of a small
number of control points, TLS datasets consist in
millions of points with a centimetric accuracy. This
complete coverage of the surface has allowed for the
morphological characterization of the rock face, the
location and volume of rockfalls (e.g. Lim et al.,
2006; Rosser et al, 2007) and the monitoring of a
likely rock slab failure.

Some limitations were also encountered in our
research: (a) Range measurements are erroneous in
the presence of vegetation and water seepage. These
measurements were not considered; (b) The
instrumental and methodological errors could be
higher than the real displacement of the probable
slab failure (see previous section); (c) The exact date
of each rockfall event is unknown because of
discrete TLS measurements; (d) A larger monitoring
period is needed to obtain a more accurate record of
large scale events; (e) A continuous, real-time TLS
record would have enabled us to better understand
the triggering factors, e.g. climatic variables.


4.4.2 Critical review of the results
The results obtained by the long-term and short-term
(i.e. monitoring by TLS) approaches are discussed
below:

Table 4.3a shows the rockfall activity for the
different types of mass movements using a long-
term approach (section 4.1.2). As regards the
number of rockfalls, a similar number of small,
medium and large number of events were recorded,
suggesting that the frequency is not dependent on
the magnitude, which appears to be inconsistent with
the inverse power law obtained in many studies (e.g.
Hungr et al., 1999; Malamud et al., 2004; Lim et al.,
2009). Furthermore, a period of recurrence of 17
years for the detachment of a single column does not
tally with our TLS results.

Table 4.3b shows the rockfall activity for the
different types of mass movements based on the
sequential comparison of TLS datasets in the
monitoring period. In contrast to table 4.3, a scale
dependency of the number of events was obtained.
However, this approach suffers from an absence of
time span, especially for large scale rockfalls.

Long-term and TLS approaches suffered from
systematic errors in the estimation of rockfall
frequency. On the one hand, the long-term approach
(table 4.3a) suffered from a bias in the recording of
the phenomena: only rockfalls with a volume above
a certain value are normally recorded and/or
detected by witnesses (e.g. Hungr et al., 1999); On
the other hand, the results of the TLS campaign
(Table 4.3b) were temporally biased: the low
frequency of medium and large scale rockfalls
demanded a longer period of study.


4.4.3 Combination of long-term and
short-term approaches

Although the limitations discussed above, an attempt
to quantify the recurrence and volume of the
rockfalls in the last 50 years is discussed as follows
(see Table 4.4): (a) the recurrence of the small scale
rockfalls was assessed using a short-term approach,
i.e. the TLS campaign; (b) the recurrence of medium
scale rockfalls was estimated as a combination of
long-term / short-term approaches. Its value range
from 2 to 11 years: on the one hand, one event was
recorded in the monitoring period, i.e. a period of
recurrence of 1.8 years; on the other hand, three
events were recorded in 1976-2008, i.e. a period of
recurrence of 10.7 years; (c) the recurrence of the
large scale rockfalls was assessed based on the long-
term approach. Two events were recorded in the last
50 years, i.e. a recurrence of 25 years. An error
margin should be considered for the following
reasons: (a) rockfall is not a homogeneous and
continuous process; (b) the results are biased in the
long-term approach (small scale rockfalls); and (c)
the results are biased in the TLS approach (large
scale events).





Abellán et al. (submitted) Chapter 4: Rockfall monitoring at Castellfollit de la Roca



Antonio Abellán Fernández Page 69

The average cliff retreat rate (e.g. Rabatel et al.,
2008; Lim et al., 2009) was estimated from 6 to 11
mm/year, assuming a detachment of material from
46 to 91.5 m
3
/year (table 4.4) and an area of the rock
face of ~8.000 m2.


4.4.4 Implications of the results
Which phenomenon represents the highest hazard?
On the one hand, small scale rockfalls account for
87 - 96 % of the events (table 4.4). However, their
effect on the cliff retreat is much lower, only from 4
to 6 %. On the other hand, large scale rockfalls make
up ~1% of the events, their effect on the cliff retreat
being much higher (from 58 to 87 %).

Which phenomenon represents the highest risk to
the population? There are two vulnerable scenarios
in the study area: the lower and the upper parts of
the cliff. No permanent vulnerable elements are
located at the base of the cliff: hence, the
detachment and propagation of the low magnitude
and high frequency rockfalls constitute a reduced
risk in this area. By contrast, the houses located on
the edge of the cliff are more likely to be damaged
by large scale rockfalls, as in the event of September
1976 (table 4.2). Subsequent efforts should be
focused on the early detection of the most hazardous,
i.e. events of great magnitude. A deformation of a
few centimetres prior to the occurrence of the April
2007 event was observed (Abellán et al., 2009).
Assuming that the most hazardous types of mass

(a) LONG-TERM APPROACH
MAGNITUDE
RECORDED
ROCKFALLS
(1)

ESTIMATED
RECURRENCE
VOLUME
> 1.5 m
3
3 43 % ~ 10 yr
(2)
4.5 m
3 (4)
0.1 m
3
/yr
(5)
0.3 %
1.5 - 150 m
3
2 28.5 % ~ 15 yr
(2)
55 m
3 (4)
1.8 m
3
/yr
(5)
3.6 %
> 150 m
3
2 28.5 % ~ 25 yr
(3)
2460 m
3 (4)
49 m
3
/yr
(6)
96.1 %
TOTAL 7 100 % - 2520 m
3
50 m
3
/yr 100 %
(1)
Number of rockfalls from the historical record (see table 4.2).
(2)
Time span of 30 years (historical record, table 4.2) divided
by the number of rockfalls.
(3)
Time span of 50 years (historical record and surveys to population, see section 4.1.2) divided by
the number of rockfalls.
(4)
Volume was estimated using the historical record (see table 4.2);
(5)
Calculated as the total volume
divided by a time span of 30 years.
(6)
Calculated as the total volume divided by a time span of 50 years.




(b) SHORT-TERM APPROACH

MAGNITUDE
RECORDED
ROCKFALLS
(1)

ESTIMATED
RECURRENCE
(2)

VOLUME
> 1.5 m
3
6 86 % ~ 0.3 yr 6 m
3 (3)
3.3 m
3
/yr
(4)
11 %
1.5 - 150 m
3
1 14 % ~ 1.8 yr 50 m
3 (3)
28 m
3
/yr
(4)
89 %
> 150 m
3
0 0 % - 0 m
3 (3)
0 m
3
/yr
(4)
0 %
TOTAL 7 100 % - 56 m
3 (3)
31 m
3
/yr
(4)
100 %
(1)
Rockfalls recorded using a sequential comparison of TLS datasets;
(2)
Time span (1.8 years) divided by the number of
rockfalls;
(3)
Volume was calculated comparing TLS datasets;
(4)
Calculated as the total volume divided by a time span of 1.8
years.

Table 4.3 (a) Long-term results based on the historical record and population surveys. Time span: 50 years prior to this study; (b)
Short-term results based on a sequential comparison of TLS datasets. Time span: 22 months.








SECTION II : Change detection Publication B: Nat. Hazards Earth Syst. Sci.


Page 70 PhD thesis

movements in the study area, i.e. medium and large
scale rockfalls, are usually preceded by slow displa-
cements (e.g. Zvelebill and Moser, 2001; Rose and
Hungr, 2007; Abellán et al., accepted), a TLS
monitoring system could be used for the early
detection of rockfalls. Further research may provide
a more accurate estimation of the rate of cliff retreat
by increasing the TLS monitoring period.



4.5. CONCLUSIONS

High resolution and high accuracy TLS datasets help
us deepen our understanding of rockfall phenomena
at Castellfollit de la Roca, especially with respect to
rockfall frequencies and magnitudes. Two types of
mass movement were detected in the monitoring
period: (a) detachment of six basaltic columns, with
a magnitude below 1.5 m
3
and (b) detachment of a
group of columns, with a magnitudes of 50 m
3
.
Moreover, the displacements of a likely slab failure
in the central part of the cliff were measured,
suggesting an apparent stationary stage
(displacements lower than 1.64 cm/year). Long-term
and TLS approaches suffered from a bias in the
recording of the rockfall phenomena. A combination
of these approaches allowed us to better estimate (a)
the magnitude and frequency of the rockfalls in the
study area and (b) the mean annual rate of retreat.
This procedure proves to be a valuable tool for
rockfall hazard assessment.



ACKNOWLEDGEMENTS

The financial support of the Spanish Ministry of
Science and Education (pre-doctoral grant 2004-
1852) is gratefully acknowledged. This work was
funded by the Natural Park of the Garrotxa Volcanic
Field, the Geomodels Institute and following
projects: MEC CGL2006-06596 (DALMASA) and
TopoIberia CSD2006-0004/Consolider-Ingenio2010.
George von Knorring improved the English draft of
the manuscript. We are very grateful to our
colleagues from the PNZVG. The valuable
comments of Guillem Gisbert and Giorgi
Khazaradze are also acknowledged. The picture of
Fig. 4.2b was taken by Llorenç Planagumà. Daily
rainfall was provided by the Servei Metereològic de
Catalunya (www.meteo.cat).


REFERENCES

Abellán, A., Vilaplana, J.M., Martínez, J., 2006. Application
of a long-range Terrestrial Laser Scanner to a detailed
rockfall study at Vall de Núria (Eastern Pyrenees, Spain).
Engineering Geology 88, 136–148.
MAGNITUDE
ESTIMATED
RECURRENCE
ESTIMATED NUM. OF
ROCKFALLS (50 YEARS)
(4)

ESTIMATED VOLUME (50 YEARS)
< 1.5 m
3
0.25 - 0.35 yr
(1)
140 - 200 87 - 96 % 135 - 200 m
3 (5)
2.5 - 4 m
3
/yr
(8)
4 – 6 %
1.5 - 150 m
3
2 - 11 yr
(2)
5 - 25 3.3 – 11 % 175 - 1375 m
3

(6)
3.5 - 27.5 m
3
/yr
(8)

5 - 40
%
> 150 m
3
20 – 30 yr
(3)
1 - 3 0.7 - 1.3 % 2000 - 3000 m
3

(7)
40 - 60 m
3
/yr
(8)

58 – 87
%
TOTAL - 146 - 228 100% 2300 - 4575 m
3
46-91.5 m
3
/yr 100%

(1)
Assessed based on the results of the TLS monitoring campaign (see table 3b) assuming an error margin of ± 20 %;
(2)

Estimated as a combination of TLS and historical record (see text);
(3)
Assessed based on the long-term approach (see table
3a) assuming an error margin of ± 20 %;
(4)
Calculated as the recurrence multiplied by a time span of 50 years;
(5)
Calculated
assuming a constant volume of 1 m
3
;
(6)
Estimated as a combination of TLS and historical record: the cumulated volume in a
time span of 30 years was calculated as 105 cubic meters, i.e. 175 cubic meters in 50 years;
(7)
Cumulated volume of
February and September 1976 events (table 2).
(8)
Calculated as the total volume divided by a time span of 50 years.


Table 4.4. Combination of short and long-term approaches: TLS monitoring campaign, historical record and population surveys.
This table attempts to quantify the recurrence and volume of the failures rockfalls in the last 50 years. The calculation of a
“probability of occurrence” is beyond the scope of this research.








Abellán et al. (submitted) Chapter 4: Rockfall monitoring at Castellfollit de la Roca



Antonio Abellán Fernández Page 71

Abellán, A., Jaboyedoff, M., Oppikofer, T., Vilaplana, J. M.,
2009. Detection of millimetric deformation using a
terrestrial laser scanner: experiment and application to a
rockfall event, Nat. Hazards Earth Syst. Sci., 9, 365-372.
Abellán, A., Vilaplana, J.M. Calvet, J., Blanchard, J.,
accepted. Rock face monitoring by Terrestrial Laser
Scanning: change detection and spatial prediction of
rockfall. Geomorphology.
Asensio, E., Vilaplana, J.M. y Abellán, A. (2009): Anàlisi
detallada de la susceptibilitat als despreniments a la
cinglera de Castellfollit de la Roca. En: Libro de actas de
las Jornades de Recerca i Gestio del Patrimoni Natural:
La vulcanología al Parc Natural de la Zona Volcánica de
la Garrotxa, Olot.
ASPRS LiDAR Committee, 2004. ASPRS Guidelines
Vertical Accuracy Reporting for LiDAR Data. Available
at <http://www.asprs.org/society/committees/lidar/
Downloads/Vertical_Accuracy_Reporting_for_Lidar_Da
ta.pdf>.
Bauer, A, Paar, G., Kaltenböck, A., 2005. Mass movement
monitoring using terrestrial laser scanner for rock fall
management, Proceedings of the First International
Symposium on Geo-Information for Disaster
Management, pp. 393–406 Delft, The Netherlands.
Biasion, A., Bornaz, L., Rinaudo, F., 2005. Laser Scanning
Applications on Disaster Management, in: Geo-
information for Disaster Management, edited by: van
Oosterom, P., Zlatanova, S., and Fendel, E. M., Springer,
Berlin, 19–33.
Chen, Y., Medioni, G., 1992. Object Modelling by
Registration of Multiple Range Images. Image and
Vision Computing 10, 145–155.
Copons, R., Vilaplana, J.M., 2008. Rockfall susceptibility
zoning at a large scale: From geomorphological
inventory to preliminary land use planning. Engineering
Geology Vol 102, Issues 3-4, 42-51.
Crosta, G.B., Agliardi, F., 2004. Failure forecast for large
rock slides by surface displacement measurements.
Canadian Geotechnical Journal, 40(1), 176-191.
Culebras, J., 2002. Castellfollit de la Roca. Quaderns de la
Revista de Girona. Diputació de Girona. 96pp.
Dewez, T., Gebrayel, D., Lhomme, D., Robin, Y., 2009.
Quantifying morphological changes of sandy coasts by
photogrammetry and cliff coasts by lasergrammetry. La
Houille Blanche (1) 32-37 DOI: 10.1051/lhb:2009002
Donville, B., 1973 Geologie néogéne et ages des eruptions
volcaniques de la Catalogne Orientale. Thèse Univ. Paul
Sabatier, Toulousse 356pp.
Dunning, S.A., Massey, C.R. Rosser, N.J., 2009. Structural
and geomorphological controls on landslides in the
Bhutan Himalaya using Terrestrial Laser Scanning.
Geomorphology, Vol. 103, Issue 1, 17-29.
Hungr, O, Evans S.G., Hazzard J., 1999. Magnitude and
frequency of rock falls and rock slides along the main
transportation corridors of southwestern British
Columbia. Canadian Geotechnical Journal 36: 224–238.
Institut Cartogràfic de Catalunya (ICC), Institut Geològic de
Catalunya (IGC) and Parc Natural de la Zona Volcànica
de la Garrotxa (PNZVG), 2007. Carta vulcanològica de
la zona volcànica de la Garrotxa [cartographic material].
Scale: 1:25.000. 93 x 96 cm. Edited by ICC.
Jaboyedoff M., Metzger R., Oppikofer T., Couture R.,
Derron M.-H., Locat J., Turmel D., 2007. New insight
techniques to analyze rock-slope relief using DEM and
3D-imaging cloud points: COLTOP-3D software. In:
Eberhardt, E., Stead, D., Morrison T.: Rock mechanics:
Meeting Society’s Challenges and demands (Vol. 2),
Taylor & Francis, 61-68.
Jaszczak, P., 2006. Terrestrial Laser Scanning for landslides
deformation monitoring. MsC thesis. Institute of
Geomatics. 39pp.
Lichti D.D., Jamtsho M., 2006. Angular resolution of
terrestrial laser scanners. Photogrammetric Record
21(114), 141–160.
Lichti, D.D., 2007. Error modelling, calibration and analysis
of an AM-CW terrestrial laser scanner system, ISPRS
Journal of Photogrammetry and Remote Sensing 61 (5),
307–324.
Lim, M., Petley, D.N., Rosser, N.J., Allison, R.J., Long, A.
J., Pybus, D., 2006. Combined digital photogrammetry
and time-of-flight laser scanning for monitoring cliff
evolution. Photogrammetric Record, 20(1), 109-129.
Lim, M., Rosser, N.J., Allison, R.J., Petley, D.N., 2009.
Erosional processes in the hard rock coastal cliffs at
Staithes, North Yorkshire, Geomorphology, doi:10.1016/
j.geomorph.2009.02.011.
Malamud B.D., Turcotte D.L., Guzzetti F., Reichenbach P.
(2004) Landslide inventories and their statistical
properties. Earth Surface Processes and Landforms, Vol.
29, n. 6, 687–711, DOI: 10.1002/esp.1064.
Mallarach, J. M., Mirabell, E, 1976. Informe Geològic sobre
les esllavissades de la tardor de 1976 a la cinglera
basàltica de Castellfollit de la Roca, La Garrotxa.
(unpublished material).
Mallarach, J.M., Riera, M., 1981. Els volcans olotins i el seu
paisatge. 62-69 i 192- 200pp. Ed. Serpa. Girona. 250pp.
Martí, J., Pujades, A., Ferrés, D., Planagumà, L., Mallarach,
J. M., 2000. El Vulcanisme. Guia de camp de la zona
volcànica de la Garrotxa. Guies del Parc Natural de la
Zona Volcànica de la Garrotxa, Departament de Medi
Ambient de la Generalitat de Catalunya, Barcelona. 106
pp.
Mascort, D., Martorell, E., Rovira, A., Montserrat, A., Pujol,
D., 2004. Estudi geologic de la zona de Castellfollit de la
Roca. Parc Natural de la Zona Volcánica de la Garrotxa,
30p.
Monserrat, O., Crosetto, M., 2008. Deformation
measurement using terrestrial laser scanning data and
least squares 3D surface matching. ISPRS Journal of
Photogrammetry and Remote Sensing, 63(1), 142-154.
Olivera, C., Redondo, E., Lambert, J., Riera Melis, A., Roca,
A., 2006. Els Terratrèmols dels segles XIV i XV a





SECTION II : Change detection Publication B: Nat. Hazards Earth Syst. Sci.


Page 72 PhD thesis

Catalunya. Generalitat de Catalunya, Institut Cartogràfic
de Catalunya, 407p.
Oppikofer, T., Jaboyedoff, M., Keusen, H.-R., 2008.
Collapse of the eastern Eiger flank in the Swiss Alps.
Nature Geosciences 1(8), 531-535.
Palau, A., 1976. Dictamen sobre las causas y remedios de las
grietas y desprendimientos en el cingle o acantilado
basaltico de Castellfollit de la Roca. Excmo.
ayuntamiento de Castellfollit de la Roca. 35pp. Girona.
Pallí, L., Trilla, J., 1976. Estudio de los desprendimientos en
el acantilado o “cingle”de Castellfollit de la Roca.
Excma. diputación provincial de Girona. 37pp. Girona.
Petrie and Toth, 2008. Introduction to laser ranging,
profiling, and scanning. In Topographic Laser Ranging
and Scanning: principles and processing. Edited by: Shan,
J., Toth, C. K., CRC Press, Taylor & Francis, 590p.
Rabatel, A., Deline, P., Jaillet, S., Ravanel, L. (2008), Rock
falls in high-alpine rock walls quantified by terrestrial
lidar measurements: A case study in the Mont Blanc area,
Geophys. Res. Lett., 35, L10502,
doi:10.1029/2008GL033424.
Ros, X., Palomar, J., Gaete, R., 1996. Estudi geotècnic del
cingle de Castellfollit de la Roca. Parc Natural de la
Zona Volcánica de la Garrotxa (PNZVG). 157 pp.
Rose, N.D., Hungr, O., 2007. Forecasting potential rock
slope failure in open pit mines using the inverse-velocity
method. International Journal of Rock Mechanics and
Mining Sciences, 44(2), 308-320.
Rosser, N.J., Petley, D.N., Lim, M., Dunning, S. A., Allison,
R.J., 2005. Terrestrial laser scanning for monitoring the
process of hard rock coastal cliff erosion. Quarterly
Journal of Engineering Geology and Hydrogeology,
38(4), 363-375.
Rosser, N.J., Lim, N, Petley, D.N., Dunning, S., Allison,
R.J., 2007. Patterns of precursory rockfall prior to slope
failure. Journal of Geophysical Research, Vol. 112, No.
F4, doi:10.1029/2006JF000642.
Sturzenegger, M., Stead, D., 2009. Quantifying discontinuity
orientation and persistence on high mountain rock slopes
and large landslides using terrestrial remote sensing
techniques. Nat. Hazards Earth Syst. Sci., 9, 267–287.
Teza, G., Galgaro, A., Zaltron, N., Genevois, R., 2007.
Terrestrial laser scanner to detect landslide displacement
fields: a new approach. International Journal of Remote
Sensing, Volume 28, Issue 16, pages 3425 – 3446.
Varnes, D. J., 1978. Slope movement types and processes. In:
Schuster, R.L., Krizek, R.J. (Eds.), Special Report 176:
Landslides: Analysis and Control. TRB, National
Research Council, Washington, D.C., pp. 11–33.
Vilajosana, I., Suriñach, E., Abellán, A., Khazaradze, G.,
Garcia, D., Llosa, J., 2008. Rockfall induced seismic
signals: case study in Montserrat, Catalonia, Nat.
Hazards Earth Syst. Sci., 8, 805-812.
Voegtle, T., Schwab, I., Landes, T., 2008. Influences of
different materials on the measurement of a Terrestrial
Laser Scanner (TLS). Proc. of the XXI Congress, The
International Society for Photogrammetry and Remote
Sensing, ISPRS2008, Vol. XXXVII, Commission V, 3-
11 July 2008, Beijing, China, 1061-1066.
Zvelebill, J., Moser, M., 2001. Monitoring Based Time-
prediction of rock falls: Three Case-Histories. Phys.
Chem. Earth (B), Vol. 26, No. 2, 159 – 167.
































CHAPTER 5
Detection of small
displacements using a TLS

Publication C: Abellán et al., (2009). Nat. Hazards Earth Syst. Sci






























































Nat. Hazards Earth Syst. Sci., 9, 365–372, 2009
www.nathazardsearthsystsci.net/9/365/2009/
© Author(s) 2009. This work is distributed under
the Creative Commons Attribution 3.0 License.












Detection of millimetric deformation using a
terrestrial laser scanner: experiment and application
to a rockfall event
A. Abellan
1
, M. Jaboyedoff
2
, T. Oppikofer
2
, and J. M. Vilaplana
1
1
RISKNAT group & GEOMODELS Institute, Depart. of Geodynamics and Geophysics, Univ. of Barcelona, Spain
2
Institute of Geomatics and Risk Analysis (IGAR), University of Lausanne, Switzerland

Received: 21 October 2008 – Revised: 25 February 2009 – Accepted: 27 February 2009 – Published: 17 March 2009

Abstract. Terrestrial laser scanning (TLS) is one of the most promising surveying techniques for
rockslope characterization and monitoring. Landslide and rockfall movements can be detected by
means of comparison of sequential scans. One of the most pressing challenges of natural hazards is
combined temporal and spatial prediction of rockfall. An outdoor experiment was performed to
ascertain whether the TLS instrumental error is small enough to enable detection of precursory
displacements of millimetric magnitude. This consists of a known displacement of three objects
relative to a stable surface. Results show that millimetric changes cannot be detected by the analysis
of the unprocessed datasets. Displacement measurement are improved considerably by applying
Nearest Neighbour (NN) averaging, which reduces the error (1σ) up to a factor of 6. This technique
was applied to displacements prior to the April 2007 rockfall event at Castellfollit de la Roca, Spain.
The maximum precursory displacement measured was 45 mm, approximately 2.5 times the standard
deviation of the model comparison, hampering the distinction between actual displacement and
instrumental error using conventional methodologies. Encouragingly, the precursory displacement
was clearly detected by applying the NN averaging method. These results show that millimetric
displacements prior to failure can be detected using TLS.


Keywords: Terrestrial Laser Scanning; rockfall; instrumental error; precursory indicator; surface
displacement














Abellán et al. (2009) Chapter 5: Detection of small displacements



Antonio Abellán Fernández Page 77

5.1 INTRODUCTION

Terrestrial laser scanning (TLS) is one of the most
promising surveying techniques for rockslope
characterization and monitoring (Bitelli et al., 2004;
Biasion et al., 2005; Abellán et al., 2006). TLS
acquires a high-resolution point cloud of the survey
scene based on the measurement of the time-of-flight
of an infrared pulse emitted in a known direction
(Slob and Hack, 2004). Three dimensional variations
of the terrain (involving landslide and rockfall
movements) can be detected by means of
comparison of sequential terrestrial laser scans
(Bauer et al., 2005; Rosser et al., 2005; Oppikofer et
al., 2008a).

One of the present challenges in rockfall hazard is
combined temporal and spatial prediction of
rockfalls. An important advance in the former is the
apparent consistency in the tertiary creep stage
(Terzaghi, 1950) of brittle failure: an acceleration
of the displacement rates prior to a failure. Current
works on failure forecasting are mainly based on
establishing inverse velocity against time
relationships (Saito, 1969; Fukuzono, 1985). The
pre-failure deformation of monitored rockfalls
ranges from a few centimetres to several decimetres,
proportional to event size (Zvelebil and Moser,
2001; Crosta and Agliardi, 2003; Rose and Hungr,
2007). One of the limitations of these works is that
the rates of displacement were acquired on single
points only (i.e. extensometers, GPS nodes and/or
total stations). The measurement of precursory
displacements over great extensions seems only to
be possible with new remote sensing techniques, i.e.
Ground- Based Interferometric Synthetic Aperture
Radar (GB-InSAR) and/or terrestrial laser scanning
(TLS).

Despite these advances in rockfall forecasting,
spatial prediction of future rockfalls over wide areas
is still unfeasible. Current efforts in source area
determination are still based on well known
techniques like the study of historical records (Ibsen
and Brunsden, 1996), statistics (i.e. multivariate
analysis, Carrara, 1983) or heuristic techniques (i.e.
expert criteria, Soeters and Van Westen, 1996). A
pioneering study that detects precursory patterns in
wide areas using TLS was made by Rosser et al.
(2007), who observed a precursory pattern of small
rockfalls leading to larger failures.

One question that remains to be resolved is as
follows: is the instrumental error of TLS systems
small enough to detect pre-failure deformation on
rockslope surfaces? Under normal conditions, the
TLS model comparison error is of the order of
centimetres (Fiani and Siani, 2005; Teza et al,
2007). Under these circumstances, the instrumental
error may mask precursory deformation. This is due
to the fact that conventional comparisons do not
fully exploit one of the main advantages of TLS:
the high quantity and density of measurements
(Monserrat and Crosetto, 2008). As a consequence,
errors in comparison can potentially be reduced by
using the information of the neighbouring points, i.e.
by filtering or interpolation (Lindenbergh and
Pfeifer, 2005).

The aims of this study are: (i) to develop a
methodology that is able to detect millimetric /
centimetric scale deformation on rockslopes using
TLS and (ii) to apply this methodology to detect a
precursory deformation on a real falling slope. An
outdoor experiment with controlled conditions of
range and deformation was performed in order to
simulate small scale deformation prior to failure.
Two different data analysis techniques were
employed: the analysis of the original, unprocessed
datasets (referred to here after as RAW data) and a
filtering technique based on the average value of the
Nearest Neighbours. Finally, these techniques were
applied to the detection of the precursory
deformation of a 50m
3
rockfall on a basalt rockface
at Castellfollit de la Roca, Catalonia, Spain.



5.2 MATERIAL AND METHODS

5.2.1 Instrument characteristics
The terrestrial laser scanning system used is an
Optech ILRIS3D, which consists of a
transmitter/receiver of infrared laser pulses
(1535 nm wavelength) and scanning optics.
Distance measurement (ρ) is based on the time-of-
flight (∆t) of the laser pulse to travel and reflect
from the surface of interest (equation 1):

ρ = c·∆t/2 [eq. 1]

where c = speed of light.

Location of each point is acquired in a polar
coordinate system (ρ, ϑ, ϕ). The horizontal and
vertical angles (ϑ and ϕ, respectively) are modified
by the scanner device using an internal system of
rotating mirrors. In our study, these angles are





SECTION II : Change detection Publication C: Nat. Hazards Earth Syst. Sci., 9, 365–372, 2009


Page 78 PhD thesis

transformed into a Cartesian coordinate system (x, y,
z) (equation 2):

[x y z]
t
= ρ [cos θ cos ϕ, cos θ sen ϕ, sen ϕ]
t

[eq. 2]

The reflectivity, i.e. the amount of reflected signal
with respect to the emitted one, is also recorded for
each point. It mainly depends on the range, angle of
incidence, material moisture and object material.
Compared to conventional surveying methods, a
TLS shows a very high data acquisition speed (up
to 2,500 points/second). Technical characteristics of
the ILRIS3D supplied by the manufacturer show
high maximum range (up to 700 m for natural
slopes) and point accuracy of 7 mm at 50m. The
object surface orientation influences the accuracy:
as the beam footprint becomes increasingly
elongate, the error is increased (Ingensand et al.,
2006). A brief discussion on TLS principles and
performances is beyond this paper, but can be found
in Teza et al. (2007).



5.2.2 Experimental setup
The experiment was performed at the Lausanne
University campus. It consists of a simulation of a
pre-failure deformation on a rockslope by a
simulated displacement of three objects, (i) a plane,
(ii) a hemisphere and (iii) an irregular form, relative
to a fixed, stable and vertical plane (Fig. 5.1). The
displacements of the three objects range between 5
and 25 mm, with an increment of 5 mm between
each scan.

After each induced displacement, a TLS point cloud
was acquired (referred to here after as data point
cloud) and compared with the initial point cloud
captured at 0 mm displacement (referred to here
after as reference point cloud). The objects were
scanned at 35'500 points/m
2
, or in terms of a square
grid at 1 point every 5.3 mm at a distance of 50 m
between the TLS and the fixed plane.

The real displacement was assessed with callipers.
The repeatability of the calliper measurements is
~0.1 mm and given that this value is approximately
2 orders of magnitude more accurate than the TLS
instrumental error, the calliper value can be
considered hereafter as the real displacement value.

Fig. 5.1 (a) General overview of the experimental setup (range = 50 m) (b) Zoom of the scanned area showing the 3 moving
objects on the fixed part. (c) Perspective views of the TLS point cloud.







Abellán et al. (2009) Chapter 5: Detection of small displacements



Antonio Abellán Fernández Page 79

5.2.3 Displacement computation
The TLS displacement values were calculated for
each object using two different approaches: (a)
RAW data, and (b) average of the 24 Nearest
Neighbours.

The RAW data displacements between the
reference and data point clouds were computed in
Innovmetrics PolyWorks™ v.9 software using a
conventional methodology (data vs. reference
comparison). Difference is therefore calculated as
normal to the stable base plane and direction of
displacement. For each object, the average
displacement (µ) and standard deviation of
displacement measures (σ) are calculated.

In order to reduce the error in RAW data
comparison, a Nearest Neighbour (NN) averaging
technique was applied. This technique consists of a
(i) data interpolation to a square grid, (ii) a search
for the k surrounding points (Davis, 1975), and (iii)
the calculation of the average value of the NN for
each point, excluding the edges. In order to obtain a
good agreement between accuracy and resolution,
different k values were tested in the experimental
case study. Algorithms that involved low numbers
of NN (k=8, 3x3 NN) retained significant noise. By
contrast, algorithms that involved a larger number
of NN (k=35, 6x6 NN) masked local scale
displacements. A k value of 24 (5×5 NN) was
selected as an optimal compromise.

Error in comparison of the sequential scans is a
function of the instrumental error, alignment error
and modelling error (Teza et al., 2007). Given that
the TLS and the object remained at the same
position during the whole experiment, alignment
error is negligible. Moreover, modelling error was
minimized using a consistent geometry and a very
large number of points. The TLS instrumental error
was calculated as the standard deviation of the
distance between the points of the fixed plane and
the best fit plane to these points (1σ = 7.2 mm,
Table 5.1).



5.3 RESULTS

5.3.1 Displacements based on RAW
data
The average displacements obtained from RAW
data comparisons between the reference and data
point clouds are shown in Table 5.1. Figure 5.2a
displays the scatter of binned RAW data
displacement values of the plane, for each
displacement step (5 to 25 mm by 5 mm
increments). A fitted Gaussian distribution is also
provided. Figure 5.2b displays RAW data vs. real
value measurements. The three objects (plane,
sphere and irregular form) show significant linear
correlation between the actual and the average TLS
value (R
2
> 0.99).

Both figures show that the range of the values of
the RAW data (7.4 mm < σ < 11.2 mm) is of the
same order of magnitude as the first and second
deformation increments of the experiment (5 and
10 mm, respectively). As a result, the real
displacement is masked by this high variability.

Figure 5.3 shows a comparison between reference
and data point clouds for displacements ranging
from 5 to 25 mm. As stated before, the real
displacements in Figures 5.3a and 5.3b (5 and
10 mm, respectively) are masked by the scattering
of the RAW data. In these figures it is difficult to
differentiate moving and fixed parts. However, the
induced displacements can be detected in Figure
5.3c (15 mm), and are clearly visible in Figures
5.3d and 5.3e (20 and 25 mm, respectively).

Summarizing, the value above which precursory
displacements can be detected (referred to here after
as threshold value) was set at 15 mm. This value
corresponds to twice the standard deviation of the
model comparison (equation 3).

Threshold value = 2 · σ
model comparison
[equation 3]


5.3.2 Nearest Neighbour averaging
Figure 5.4 shows the comparisons between
reference and data point clouds for displacements
ranging from 5 to 25 mm using the 5x5 nearest
neighbour averaging algorithm. As the
instrumental error is filtered out using this method,
the real displacement is visible in Figures 5.4a to
5.4e (5, 10, 15, 20 and 25 mm, respectively).

In contrast to the low precision (or high variability)
obtained using conventional methodologies (Fig.
5.3), the displacement was more accurately
computed after NN averaging, even for the smallest
displacement value (5 mm, Fig. 5.4a). The standard
deviation of the model comparison is 1.3 mm, six
times more accurate using NN averaging rather than





SECTION II : Change detection Publication C: Nat. Hazards Earth Syst. Sci., 9, 365–372, 2009


Page 80 PhD thesis






(a) TLS
Measurements (mm)

(b) Calliper Measurements
(mm)

variable
Num.
of
points
Mean
value
(µ)
St.
Dev.
(σ)
Min.
value
Max.
value
Mean
displa-
cement

a

b

c

d
Mean
displa-
cement
Fixed part
43,561 -0.31 7.21 -36.19 32.20 0.00 - - - - -
Plane
6,971 -0.11 7.26 -41.46 32.45 0.20 17.07 17.10 19.12 17.70 0.00
Hemisphere
6,175 -0.62 9.67 -78.76 71.01 -0.31 18.63 18.23 18.80 18.97 0.00
0
m
m

Irregular
20,049 -0.37 11.46 -234.16 195.79 -0.06 19.20 18.85 19.57 18.68 0.00
Fixed part
75,741 -0.06 7.28 -27.69 27.51 0.00 - - - - -
Plane
6,670 4.48 7.19 -21.90 32.01 4.53 21.95 22.07 24.10 23.05 5.05
Hemisphere
6,245 5.52 10.02 -74.84 87.34 5.57 23.40 23.25 23.68 24.05 4.94
5
m
m

Irregular
19,364 4.67 11.56 -207.78 229.83 4.73 24.08 23.65 24.95 23.24 4.91
Fixed part
48,129 -0.47 7.21 -40.45 46.57 0.00 - - - - -
Plane
6,943 9.18 7.40 -21.82 38.30 9.65 26.90 22.27 29.03 27.47 8.67
Hemisphere
6,101 9.83 9.57 -49.09 50.67 10.30 28.45 28.45 28.70 29.08 10.01
1
0
m
m

Irregular
20,031 9.20 11.09 -209.81 229.71 9.67 28.85 28.83 29.55 27.60 9.63
Fixed part
47,513 -0.41 7.15 -40.05 33.84 0.00 - - - - -
Plane
6,823 14.04 7.54 -13.76 42.15 14.45 30.60 31.98 33.74 31.25 14.14
Hemisphere
6,114 14.90 10.01 -72.53 57.90 15.31 33.30 33.18 33.38 33.60 14.71
1
5
m
m

Irregular
19,444 14.47 11.79 -206.78 246.84 14.88 33.60 33.68 34.38 33.57 14.73
Fixed part
44,608 0.22 7.10 -38.72 36.74 0.00 - - - - -
Plane
6,990 19.79 7.38 -11.67 50.24 19.58 37.68 37.13 39.23 36.38 19.85
Hemisphere
6,036 20.30 10.44 -63.85 103.17 20.09 38.65 37.75 38.68 39.05 19.88
2
0
m
m

Irregular
19,084 19.57 11.85 -209.06 237.96 19.35 38.78 37.10 39.58 38.04 19.30
Fixed part
43,825 -0.04 7.13 -41.24 29.79 0.00 - - - - -
Plane
6,903 24.46 7.48 -2.56 54.67 24.51 41.40 41.93 43.90 41.58 24.45
Hemisphere
6,108 25.11 10.69 -58.36 102.49 25.16 43.75 42.90 43.43 44.08 24.88
2
5
m
m

Irregular
19,395 24.82 12.09 -200.08 243.48 24.86 43.48 42.10 44.78 42.97 24.26

Table 5.1. (a) TLS measurements. (b) Calliper measurements.








Abellán et al. (2009) Chapter 5: Detection of small displacements



Antonio Abellán Fernández Page 81











Fig 5.2. (a) Normalised number of observations (Y axis) for each interval (0.5mm classes) of the RAW data (plane object).
Curves show fitted Gaussian distributions (µ = mean value; σ = standard deviation) for 5, 10, 15, 20 and 25 mm (from left to
right, respectively). (b) Mean value of the RAW data displacements vs. real value of the displacement (calliper) for the three
objects (plane, hemisphere and a irregular form). The error bars represent 1σ standard deviation (7.4, 10.1 and 11.2 mm
respectively). Standard deviation for the plane is significantly lower than for the hemisphere and the irregular form.











SECTION II : Change detection Publication C: Nat. Hazards Earth Syst. Sci., 9, 365–372, 2009


Page 82 PhD thesis











Fig. 5.3. RAW data displacement measurements (7.4mm< σ <11.2mm) for different induced displacements: (a) 5 mm, (b) 10 mm,
(c) 15 mm, (d) 20 mm, (e) 25 mm. The colour scale (in millimeters) indicates displacements of up to 30mm.


Fig. 5.4. Displacement measurements after the 5×5 closest neighbours averaging for different induced displacements (σ
plane
=
1.3mm): (a) 5 mm, (b) 10 mm, (c) 15 mm, (d) 20 mm, (e) 25 mm. The colour scale (in millimeters) indicates displacements of up
to 30mm.






Abellán et al. (2009) Chapter 5: Detection of small displacements



Antonio Abellán Fernández Page 83

conventional methodologies. Using equation 3, the
threshold value using NN technique was potentially
set at 2.6 mm.




5.4 Application to a rockfall
event

The basalt cliff at Castellfollit de la Roca, Catalonia,
Spain (Fig. 5.5a) has been monitored using TLS by
the RISKNAT group since March 2006. The
research is focused on quantifying the volume and
frequency of current failures in order to estimate
future rockfall hazard concern surrounds houses
located at the edge of the rockface that could be
affected by cliff retreat. The historical rockfall
record at this site shows: (i) few rockfall events
(~1 m
3
) per year; (ii) medium scale rockfalls (~50
to 250 m
3
) with a period of recurrence of about 10
years; (iii) larger scale rockfalls (~1,000 to
2,500 m
3
) with a recurrence of 50 years (Abellán et
al, 2008). A set of 50 m
3
columnar basalt blocks fell
(Fig. 5.5b), after a period of continuous rainfall
(100 mm in 1 week) in April 2007, which we
analyze below.

TLS point clouds were acquired using an Optech
ILRIS3D TLS (i) in September 2006, (ii) in
December 2006, (iii) a few days before and (iv) a
few days after the April 2007 failure. The scans
were performed from the opposite side of the valley
at a mean range to the rockface of 190 m. The
effective resolution (mean point spacing) on the
cliff was defined as 70 mm. The (iii) point cloud
served as a reference for comparison with datasets
Fig. 5.5. (a) Study area: Basaltic cliff at Castellfollit de la Roca. The frame corresponds to Fig. 5b (b) Comparison of TLS models
(iii) and (iv) (March and April 2007, respectively) showing the 50 m
3
rockfall event in April 2007. Colour scale from 0 (blue) to
1.5 m (red).






SECTION II : Change detection Publication C: Nat. Hazards Earth Syst. Sci., 9, 365–372, 2009


Page 84 PhD thesis

(i) and (ii).

The comparison of the reference and data point
clouds using the RAW data is shown in Figure 5.6.
Figures 5.6a and 5.6b suggest that the cliff
underwent deformation in the 6 months prior to the
failure. However, owing to the high variability
discussed in sections 5.2 and 5.3, we were unable to
clearly delimit the extent of deformation. The
model comparison error as described by σ is
calculated for the stable parts of the cliff, as
18.0 mm. Equation 3 shows that precursory
displacement under 2·σ cannot be detected using
this approach. This threshold is therefore set at in
36.0 mm.

A detailed analysis of the reference and data point
clouds, using the 24 Nearest Neighbour averaging,
shows a deformation of the cliff sector (Figure 5.7).
The calculated maximum precursory displacement
was 45 mm (Figure 5.7b), a value slightly higher
than that of the threshold mentioned above. For this
reason, the attempts to isolate the precursory
displacement from the instrumental error using
conventional methodologies were unsuccessful.
Error using the NN averaging technique (1σ) over a
stable part was calculated as 6.4mm, which is 3
times more accurate than with conventional
methodologies.



5.5 DISCUSSION and CONCLUSIONS

Standard deviation of the instrumental error was
calculated as 7.2 mm at a distance of 50 m (table
5.1). This value is in the same order of magnitude
as the error in the RAW data comparisons for the
plane, the hemisphere and the irregular form (7.4,
10.1 and 11.2 mm, respectively). The RAW data
comparisons show that errors increase with the
complexity of the shape (1.0, 1.4 and 1.5 times the
standard deviation of the instrumental error,
respectively). This could be due to: a low
reflectivity, a high incidence angle and/or to a
different surface character. On the one hand, several
authors (e.g. Soudarissanane et al., 2008; Voegtle
et al., 2008) demonstrated the influence of low
reflectivity values and large incidence angles in
lowering the accuracy. Evidences of these effects
are found in the hemisphere: the low values of
reflectivity and accuracy were found around the
external part of the object (Fig.5.1 and 5.3). By
filtering these points, the quality of the overall
measurement could be enhanced. On the other hand,
the influence of the object material suggested by
Voegtle et al. (2008) was negligible in our study.

Precursory displacements lower than 2σ cannot be
detected with certainty (equation 3) using RAW
datasets, in the experimental setup, or in the real
case study. In contrast, the datasets averaged by a
Nearest Neighbour method enabled a more precise
measurement of these millimetric displacements. Its
application to the April 2007 Castellfollit de la
Roca rockfall underlines the utility of NN method
in real case studies.

New ways to fully exploit the huge quantity of
information provided by TLS point clouds are still
needed. In our experimental case study, the
direction of movement was restricted in 2
dimensions (z,x) and 3 rotations (x,y,z edges). As a
result, we defined the vector of comparison along
the permitted deformation direction (y, range).
However, in a real case study, the direction of
displacement is generally not known in advance.
This could be a limitation of the comparison
technique although the displacement direction can
be assessed defining different roto-translation (RT)
matrices over discrete parts of the slope (Monserrat
and Crosetto, 2008; Oppikofer et al., 2008b),
assuming that the nature of the distribution remains
constant. Either using the NN as in the RT
techniques, the larger the number of points involved
in the calculation, the greater the potential accuracy.

The main advantage of using TLS instead of point
based monitoring techniques is the effectively
complete measurement of the rock face. If
precursory displacement could always be detected
on a rock slope prior to a rockfall event, a fixed
TLS system collecting a continuous record of the
3D geometry of the slope could be established. For
each point on the square grid described in section
5.2.3, displacement vs. time could be plotted and
evolution of deformation monitored. The temporal
prediction of rockfalls and an early warning system
could be based on the same framework as that used
for a single point measurement, such as inverse
velocity – time relationships (Fukuzono 1985).

The NN technique require validation in more cases
to test its applicability under real conditions
involving different surface materials and different
type of failures. Future work will focus on the
detection of precursory displacements at different
ranges (i.e. from 100 up to 500 m) and for variable
displacement directions.







Abellán et al. (2009) Chapter 5: Detection of small displacements



Antonio Abellán Fernández Page 85


Fig. 5.6. RAW data displacement measurements prior to the 50m3 rockfall; (a) cumulated displacement for 3 months comparison;
(b) cumulated displacement for 6 months comparison. Precursory displacement cannot be clearly detected using this approach.
σRAW data=18.0 mm.






SECTION II : Change detection Publication C: Nat. Hazards Earth Syst. Sci., 9, 365–372, 2009


Page 86 PhD thesis


Fig. 5.7. Filtered displacement measurements prior to the 50m3 rock-fall using the NN averaging technique. (a) cumulated
displacement for 3 months comparison; (b) cumulated displacement for 6 months comparison. A centimetric precursory
displacement is observed in the middle of the figure. σNN =6.4 mm.






Abellán et al. (2009) Chapter 5: Detection of small displacements



Antonio Abellán Fernández Page 87

ACKNOWLEDGEMENT

The academic stay at IGAR-UNIL (ref. AP-2007-
1852) and a FPU pre-doctoral grant (AP-2004-1852)
was financed by the Ministry of Education of Spain .
This research was funded by Geomodels Institute,
TopoIberia CSD 2006 -0004 / Consolider - Ingenio
2010 and MEC project CGL2006 – 06596
(DALMASA). We are indebted to our colleagues of
IGAR (A. Pedrazzini) and CNRS (J. Travelletti) for
their assistance with the experimental setup. Thanks
are also due to George von Knorring for improving
the English version of the manuscript. And finally,
we are very grateful to M. –H. Derron, N.J. Rosser,
H. Hack and an anonymous referee for their critical
review of the manuscript.



REFERENCES

Abellán, A., Vilaplana, J.M. and Martínez, J.: Application of
a long-range terrestrial laser scanner to a detailed
rockfall study at Vall de Núria (Eastern Pyrenees, Spain).
Engineering geology, 88(3-4), 136-148, 2006.
Abellán, A., Vilaplana, J.M., Calvet, J. and García, D.:
Seguimiento (monitoring) de laderas rocosas mediante
un Láser Escáner Terrestre. Zona de estudio piloto en
Castellfollit de la Roca. Cataluña. Geo-temas 10, 1385-
1388, 2008.
Bauer, A., Paar, G. and Kaltenböck, A.: Mass Movement
Monitoring Using Terrestrial Laser Scanner for Rock
Fall Management. In: van Oosterom, P., Zlatanova, S.
and Fendel, E.M. (Editors), Geo-information for Disaster
Management. Springer, Berlin, 393-406, 2005.
Biasion, A., Bornaz, L. and Rinaudo, F.: Laser Scanning
Applications on Disaster Management. In: van Oosterom,
P., Zlatanova, S. and Fendel, E.M. (Editors), Geo-
information for Disaster Management. Springer, Berlin,
19-33, 2005.
Bitelli, G., Dubbini, M. and Zanutta, A.: Terrestrial laser
scanning and digital photogrammetry techniques to
monitor landslide bodies, Proceedings of the XXth
ISPRS Congress Geo-Imagery Bridging Continents,
Istanbul, Turkey, ISPRS, 246–251, 2004.
Carrara, A.: Multivariate models for landslide hazard
evaluation. Mathematical Geology, Vol. XXV: 403-426,
1983.
Crosta, G.B. and Agliardi, F.: Failure forecast for large rock
slides by surface displacement measurements. Canadian
Geotechnical Journal, 40(1), 176-191, 2003.
Davis, P. J.: Interpolation and Approximation, Dover
publications, New York, 1975.
Fiani, M. and Siani, N.: Comparison of terrestrial laser
scanners in production of DEMs for Cetara tower.
Proceedings CIPA XX International Symposium, Torino,
2005.
Fukuzono T.: A new method for predicting the failure time
of a slope. In: Proceedings of the fourth international
conference and field workshop on landslides. Tokyo:
Japan Landslide Society; p. 145–50, 1985.
Ibsen, M.L. and Brunsden, D.: The nature, use and problems
of historical archives for the temporal occurrence of
landslides, with specific reference to the south coast of
Britain, Ventnor, Isle of Wight. Geomorphology Vol. 15,
Issues 3-4: Landslides in the European Union, 241-258,
1996.
Ingensand, H. Metrological aspects in terrestrial laser-
scanning technology. In Proceedings of the 3rd IAG/12th
FIG Symposium, Baden, Austria. 2006
Lindenbergh, R. and Pfeifer, N.: A statistical deformation
analysis of two epochs of terrestrial laser data of a lock,
Proceedings of the 7th Conference on Optical 3D
Measurement Techniques, Vienna, Austria, 61-70, 2005.
Monserrat, O. and Crosetto, M.: Deformation measurement
using terrestrial laser scanning data and least squares 3D
surface matching. ISPRS Journal of Photogrammetry and
Remote Sensing, 63(1), 142-154, 2008.
Oppikofer, T., Jaboyedoff, M. and Keusen, H.-R.: Collapse
at the eastern Eiger flank in the Swiss Alps. Nature
Geosciences 1(8), pp. 531-535, 2008a.
Oppikofer, T., Jaboyedoff, M., Blikra, L. H. and Derron, M.-
H.: Characterization and monitoring of the Åknes
rockslide using terrestrial laser scanning, in: Proceedings
of the 4th Canadian Conference on Geohazards: From
Causes to Management, edited by: Locat, J., Perret, D.,
Turmel, D., Demers, D., and Leroueil, S., Presse de
l'Université Laval, Québec, Canada, 211-218, 2008b.
Rose, N.D. and Hungr, O.: Forecasting potential rock slope
failure in open pit mines using the inverse-velocity
method. International Journal of Rock Mechanics and
Mining Sciences, 44(2), 308-320, 2007.
Rosser, N.J., Petley, D.N., Lim, M., Dunning, S.A. and
Allison, R.J.: Terrestrial laser scanning for monitoring
the process of hard rock coastal cliff erosion. Quarterly
Journal of Engineering Geology and Hydrogeology,
38(4), 363-375, 2005.
Rosser, N.J., Lim, N, Petley, D.N., Dunning, S. and Allison,
R.J.: Patterns of precursory rockfall prior to slope failure.
Journal of Geophysical Research, 112(F4),
F04014(14pp.), 2007. doi:10.1029/2006JF000642
Saito, M.: Forecasting time of slope failure by tertiary creep.
In: Proc. 7th Int. Conf on Soil Mechanics and
Foundation Engineering, Mexico City, vol. 2. pp. 677–
683, 1969.
Slob, S. and Hack, R.: 3D Terrestrial Laser Scanning as a
New Field Measurement and Monitoring Technique. In:
Hack, R., Azzam, R. and Charlier, R. (Editors),
Engineering Geology for Infrastructure Planning in
Europe. A European Perspective. Lecture Notes in Earth
Sciences. Springer, Berlin / Heidelberg, 104, 179-190,
2004.





SECTION II : Change detection Publication C: Nat. Hazards Earth Syst. Sci., 9, 365–372, 2009


Page 88 PhD thesis

Soeters, R. and Van Westen, C. J.: Slope Instability
recognition, analysis and zonation. In Landslides:
Investigation and mitigation. Special report 247,
Transportation Research Board. National Research
Council, Washington D.C., pp. 129-177, 1996.
Soudarissanane, S., Lindenbergh, R. and Gorte, B. Reducing
the error in terrestrial laser scanning by optimizing the
measurement set-up. Proc. of the XXI Congress, The
International Society for Photogrammetry and Remote
Sensing, ISPRS2008, Vol. XXXVII, Commission V, 3-
11 July 2008, Beijing, China, 615-620, 2008.
Teza, G., Galgaro, A., Zaltron, N. and Genevois, R.:
Terrestrial laser scanner to detect landslide displacement
fields: a new approach. International Journal of Remote
Sensing, Volume 28, Issue 16, pages 3425 – 3446, 2007.
Terzaghi, K.: Mechanism of landslides. In Application of
geology to engineering practice (Berkeley Volume).
Edited by S. Paige. Geological Society of America,
Washington D.C., pp. 83–123, 1950.
Voegtle, T., Schwab, I. and Landes, T.: Influences of
different materials on the measurement of a Terrestrial
Laser Scanner (TLS). Proc. of the XXI Congress, The
International Society for Photogrammetry and Remote
Sensing, ISPRS2008, Vol. XXXVII, Commission V, 3-
11 July 2008, Beijing, China, 1061-1066, 2008.
Zvelebil J. and Moser, M.: Monitoring based time-prediction
of rock falls: Three Case-Histories. - Phys. Chem. Earth
(B), Vol. 26, No. 2, pp. 159 – 167, 2001.




































CHAPTER 6
Rockfall detection and
prediction using a TLS

Publication D: Abellán et al., (accepted). Geomorphology




























































Detection and spatial prediction of rockfalls by
means of terrestrial laser scanning monitoring
Antonio Abellán
a, b,*
, Joan Manuel Vilaplana
a, b
, Jaume Calvet
a, b
, Julien Blanchard
a

a
RISKNAT Research Group,. Department of Geodynamics and Geophysics, Faculty of Geology,
University of Barcelona, C/Martí i Franquès s/n. 08028 Barcelona.

b
GEOMODELS Institute, Parc Cientific de Barcelona, Universitat de Barcelona, C/Adolf Florensa s/n.
08028 Barcelona.

Received: 3 March 2009 – Refereed: 24 April 2009


Abstract. This is a case study of rock face monitoring using a ground based remote sensing
technique: Terrestrial Laser Scanner (TLS). The pilot study area corresponds to the main
scarp of an old slide located at Puigcercós (Catalonia, Spain). Four different datasets of the
slope were acquired from Sept. 2007 to July 2008 in an ongoing project. The
geomorphological evolution of the rock face, i.e. volume and frequency of rockfall, was
studied by a sequential comparison of datasets. In addition to the detection of past events, this
study deals with the spatial prediction of rockfall through the detection of the pre-failure
deformation. The high resolution, accuracy and maximum range of the TLS offer interesting
prospects for both spatial location and prediction of rockfall.



Keywords: Laser Scanning, Monitoring, Pre-failure deformation, Slope creep, Displacement
detection, Spatial prediction, Rockfall, Toppling.






















Abellán et al. (accepted) Chapter 6: Rockfall detection and prediction




Antonio Abellán Fernández Page 93

6.1 INTRODUCTION

Minor scale rockfalls (up to a hundred m
3
) are the
most frequent type of landslides on steep slopes in
mountain areas (Copons and Vilaplana, 2008),
marine cliffs (Rosser et al., 2005), etc. Since
rockfall is the fastest type of landslide (Varnes,
1978), its impact energy (and hence the geological
hazard) can reach very high values (Agliardi and
Crosta, 2003). When this natural phenomenon
interacts with populations, buildings and/or
infrastructures and when the value of the risk
exceeds that of the acceptable risk (Fell et al.,
2008), the hazard, exposure and/or vulnerability
parameters should be reduced. If corrective or
preventive measurements are not taken, rockfall
will continue to be the primary cause of landslide
fatalities in some countries (e.g. Italy; Guzzeti et al,
2000).

There are a number of approaches to the study of
rockfall phenomena such as historical inventories
(e.g. Malamud et al., 2004), susceptibility
assessment (e.g. Frattini et al., 2008), frequency
estimation (e.g. Stoffel et al., 2005) and hazard
and/or risk assessment (e.g. Copons et al., 2005). A
landslide hazard assessment should ideally
determine where and when a slope failure is likely
to occur (Guzzetti et al., 2004). As discussed in the
work of Van Westen et al. (2006), these two
questions do not have a simple answer. Both
temporal and spatial predictions are usually based
on the measurement of some indicators before
failure, e.g. crack opening, acoustic disturbances,
micro-seismicity, and pre-failure deformation.
Current works on large scale (from 10
4
to 10
6
m
3
)
landslide forecasting (Crosta and Agliardi, 2004;
Rose and Hungr, 2007) are based on the detection
of pre-failure deformation (Terzaghi, 1950; Saito,
1969; Voight, 1989). Except for the work of Rosser
et al. (2007), which is based on an increase in
rockfall activity as a precursory indicator of larger
failures, there is little bibliography on the spatial
prediction of minor scale rockfalls.

The aforementioned studies on landslide forecasting
employed point-based instruments of measurement
(e.g. extensometers and total stations) to monitor
the moving area. Despite their accuracy, these
instruments suffer from certain disadvantages, such
as the relatively low resolution and the need for a
stable point in order to compare with the moving
part of the slope. Moreover, since the moving area
is often an unknown factor, a method to detect
moving areas is still required. The possibility of
acquiring datasets with high accuracy and spatial
resolution, using laser, optical and/or radar
technologies, mounted on terrestrial, aerial and/or
satellite instrumentals are currently opening up new
ways to visualize, model and interpret surface
processes (e.g. Abellán et al., 2006; Dunning et al.,
2009). One of these new remote sensing tools, a
Terrestrial Laser Scanner (TLS) was used in this
study for the detection and spatial prediction of
rockfall at the Puigcercós test site (Catalonia, Spain).
The first part of this research deals with the
geomorphological evolution of the rock face: the
main events that occurred between Sept. 2007 and
July 2008 are described. The second part of this
study deals with the detection of the pre-failure
deformation as a precursory indicator of rockfall. In
contrast to point-based instruments, we were able to
detect this deformation over the whole rock face.
This approach could answer the important question:
where will rockfall occur?


6.1.1 Study area

The rock face studied (Fig. 6.1) is the main scar of a
landslide that took place on the night of 13 January
1881. According to Vidal (1881) who described the
event, a continuous period of rain could have been
the triggering factor. This landslide was classified
as a complex roto-translational slide (Varnes, 1978;
Corominas and Alonso, 1984), evolving into an
earth flow in its zone of accumulation. The
displaced material occupied an area of 78,520 m
2
.

The event destroyed part of the old village of
Puigcercós. The remains of some houses are still
visible on the upper part of the slope adjacent to the
scar (see Fig. 6.1). From a geological point of view,
we can observe an alternation of grey marls,
sandstones, silt and clays in the scar. More
information on the history and geometry of the
landslide can be found in Blanchard et al. (2008).

Minor rockfalls currently constitute the main
geomorphological process. The detection and
prediction of these rockfalls are the main aims of
this study. Vertical discontinuities with a strike
almost parallel to the slope face appear on the scar.
As shown in Blanchard et al. (2008), two of these
discontinuities intersect with the upper part of the
slope forming wide crown cracks. The aperture of
these discontinuities increases upwards. According
to the failure types described in Goodman and Bray
(1976), this family of discontinuities can cause a
toppling type of failure, i.e. a progressive rotation of





SECTION II: Change detection Publication D: Geomorphology


Page 94 PhD thesis

columns or blocks on a fixed base (Muller, 1968).
Sliding, excavation and erosion of the toe of the
slope are conditioning/triggering factors of the
toppling process, and this process regresses into the
rock mass with the formation of new deep tension
cracks (Wyllie et al., 2004).

The scar was selected as a pilot area for rockfall
monitoring using a TLS. The reasons for the
selection were (a) the signs of current rockfall
activity (fresh surface, scree deposits and lack of
vegetation on the toe of the slope), (b) manageable
dimensions (25 m height × 150 m width) and (c) the
favourable conditions of the area from a remote
sensing point of view (the semi-circular geometry
of the scar and gentle topography of the zone of
depletion allowed us to acquire the datasets from a
single station).



6.2. Materials and methods

6.2.1. Terrestrial Laser Scanner (TLS)
The remote sensing tool employed in this study is a
TLS. This instrument is also known as a ground-
based LIDAR (Light Detection And Ranging). We
used an Optech Intelligent Laser Ranging and
Imaging System (ILRIS3D). The instrument
consists of a transmitter/receiver of infrared laser
pulses (1535 nm wavelength) and a scanning device.
The laser beam is directly reflected on the land
surface, obviating the need for the existence of
intermediate prism reflectors. Distance
measurement (range, ρ) is based on the time-of-
flight (∆t) of the laser pulse (equation 1, Petrie and
Toth, 2008):

ρ = c·∆t/2 (1)

where c = speed of light.
.
This distance is measured for each point of area
equal to the spot dimension (S
D
). This parameter
increases its magnitude with the range owing to the
laser beam divergence process (equation 2, Petrie
and Toth, 2008):

S
D
= 0.17×10
-3
·ρ + 12 (2)

The scanning device uses an internal system of
rotating mirrors. Coordinates of each point are
acquired in a polar system (ρ: range; ϑ: horizontal
angle; ϕ: vertical angle). The binary file provided
by the TLS was transformed into a Parametric
Image Format (PIF) using Parser v 4.3.5.4,
obtaining the coordinates in a Cartesian system (x, y,
z):

[x y z]
t
= ρ [cos θ cos ϕ, cos θ sin ϕ, sin ϕ]
t
(3)

Besides the 3D coordinates, the device is able to
acquire the value of intensity (l) of reflectivity of
the land surface on which the laser beam is focused.
This parameter is defined as the amount of reflected
signal with respect to the emitted one. It mainly
depends on the range, incidence angle and object
material. Compared to conventional surveying
methods, a TLS shows a very high data acquisition
speed (up to 2,500 points s
-1
). Technical
characteristics of the ILRIS3D supplied by the
manufacturer show a high maximum range (up to
700 m for natural slopes) for a wide variety of
materials and a high point accuracy (σ ~ 0.7 cm at a
range of 100 m). Lichti (2007) demonstrated the
influence of the incidence angle on the accuracy: as
the beam footprint becomes increasingly elongate,
the error is increased. Discussion on TLS principles


Fig. 6.1. Picture of the study area showing the rock face and the upper part of the scree deposit. Ruins of the old village of
Puigcercós can be seen at the top of the slope.






Abellán et al. (accepted) Chapter 6: Rockfall detection and prediction




Antonio Abellán Fernández Page 95

and performances is beyond the scope of this paper,
but can be found in Teza et al. (2007). Given its
high accuracy and point density, the TLS
instrument is currently being used in different fields
related to natural hazards, such as the monitoring of
rockslopes (Lim et al., 2006; Oppikofer et al., 2008),
the characterization of 3D discontinuities
(Jaboyedoff et al., 2007) and landslides (Dunning et
al., 2009).


6.2.2. Data acquisition
The first dataset was acquired on September 29,
2007 (referred to here after as reference point
cloud). Data acquisition was repeated 59 days after
S
0
(period i), 135 days after S
0
(period ii) and 303
days after S
0
(period iii). These data acquisitions
(referred to here after as data point clouds) define
the three intervals considered in this paper: the 1st
interval (59 days) is between the 1st and 2nd data
acquisition; the 2nd interval (76 days) is between
the 2nd and 3rd data acquisition; and the 3rd
interval (163 days) is between the 3rd and 4th data
acquisition.

As stated above, the datasets were acquired from a
single station. Three scans were necessary to cover
the whole rock face. These scans were aligned and
merged in a single file, creating a final 3D model
(Fig. 6.2).


Fig. 6.2. Data acquisition performed with the TLS. The final 3D model of the study area was generated by aligning the
datasets. The contact between rock face and the scree deposits is indicated by a black line. (a) frontal view: “–Y axis”; (b) left
view: “-X axis”; (c) zenithal view: “+Z axis”.






SECTION II: Change detection Publication D: Geomorphology


Page 96 PhD thesis

Standard deviation (1σ) of the instrument on “real
conditions”, between two scans of the rock face
acquired consecutively, was calculated as 1.68 cm.
Data artifacts (less than 0.5% of the population)
were found in the boundaries of the occluded parts
of the slope with respect to the line-of-sight. The
mean point spacing of the acquired datasets ranged
from 4 to 6 cm. Each point of the reference point
cloud was defined as a node for a Triangle Irregular
Network (TIN) surface.From now, this This surface
will be referred to as the surface of reference (S
0
)
hereinafter.


6.2.3. Comparison of sequential scans
3D temporal variations of the terrain were detected
by comparing sequential datasets in accordance
with the methodology developed by Rosser et al.
(2005) and Lim et al., (2006). Some examples of the
use of this methodology can be found in Oppikofer
et al. (2008). A brief description of the main steps is
as follows: (i) First data acquisition (see previous
section); (ii) Construction of the “Surface of
reference” (S
0
) by aligning the first datasets; (iii)
Acquisition of a new data point cloud after a certain
interval of time; (D
1
, D
2
,….. D
n
); (iv) Alignment of
these datasets with S
0
(explained below). (v)
Comparison between S
0
and successive D
i
; and (vi)
calculation of the differences for each period of
comparison.

The roto-translation parameters of the alignment
matrix (step iv) were obtained in three phases: a
preliminary registration was performed by a visual
identification of homologous points. The alignment
was subsequently optimized using an Iterative
Closest Points (ICP) procedure (Chen and Medioni,
1992). Through this algorithm, the differences
between points were progressively reduced by a
minimization of a mean square cost function. The
final improvement was obtained by progressively
reducing the “search distance” parameter up to a
few centimetres. As a consequence, the metric scale
changes (e.g. rockfall) that occurred in the different
intervals hardly affected the global alignment,
making alignment error negligible. Apart from the
recorded metric-scale rockfalls, the surface of the
cliff remained practically unchangeable during the
scan comparison.

The single point distances between the surface of
reference and the successive data point clouds were
computed with Innovmetrics PolyWorks™ v.9
software using a conventional methodology (data vs.
reference comparison). The direction of comparison
was defined as the normal vector of the rock face at
its central part (Y axis). For comparison, the origin
of the Cartesian Coordinate System (P
0
= 0,0,0) was
set at the centre of the TLS instrument.

Differences in each point (D
i
) were calculated as:

D
i
= Distance [S
i
P
0
] - Distance[S
0
P
0
]
(3)

Part of D
i
value is due to instrumental and
methodological errors. The other part is due to a
real deformation or to a detachment of certain parts
of the slope. As regards the sign criteria, we used
positive values when the time of flight of the laser
signal for “surface of comparison” is higher than
that of “surface of reference” (see Fig. 6.3). As a
consequence, positive values equal a lack of
material at a given point, i.e. detachment of the
material. Likewise, negative values equal an
increase in material, i.e. scree deposits, or a
displacement towards the origin of coordinates. In
the second part of our study, the negative
displacements reflect the pre-failure deformation on
part of the slope. The exact date of each event is an
unknown factor because of the discrete
characteristics of the data acquisition and the lack of
eyewitnesses,.

Calculation of the volume of the main events was
made using the “surface to plane” command of the
IMInspect module (PolyWorks). In line with Rosser
et al. (2005), rockfalls with a volume under 0.001
m
3
were not considered in this study.


6.2.4. Detection of the pre-failure
deformation
The pre-failure deformation on certain parts of the
rock face (slope creep, Terzaghi, 1950) reflects the
tensional state of the slope prior to the failure. This
deformation is quantified in a discrete way through
the displacement measured at each point of the
dataset. Maruyama and Kozima (1994)
demonstrated that the amount and duration of the
slope creep has a positive relationship with the
dimensions of a soil slide. Moreover, this duration
could depend on external triggering factors, such us
rainfall and freeze-thaw. In the case of sudden
seismic events, the pre-failure deformation and the
failure may occur in a short period of time with the
result that the slope creep is masked.

As demonstrated in Abellán et al., (2009), small
displacements cannot be detected with certainty
using the original, unprocessed datasets. In this
study, the threshold from which a displacement can





Abellán et al. (accepted) Chapter 6: Rockfall detection and prediction




Antonio Abellán Fernández Page 97

be detected was defined as the double of the
standard deviation of the instrumental error. In our
research, the amount of slope creep is a few cm,
being in the same order of magnitude as the
methodological error. In order to reduce this error,
the scans were compared on a one to one basis, i.e.
avoiding the alignment of the whole surface.
Moreover, a nearest neighbour (NN) filtering
technique was applied to the original D
i
dataset,
enabling the accurate detection of these
displacements. This technique consists of the search
for the k surrounding points and the calculation of
the average value of the NN for each point. In this
case study, the search for the k = 24 closest points
was computationally optimized by using a regular
(square) mesh. The areas of the rock face that
showed a negative displacement are shown in
subsequent sections. A complete explanation of the
methodology applied to displacement detection and
its application to an experimental case is contained
in Abellán et al. (2009).




Fig. 6.3. Main events during the period of study (2007.09.29 – 2008.07.23). Events are chronologically ordered, from A to E. Sign
criteria applied to the differences calculation (Di) is showed on the colour scale (see text for a complete explanation).






SECTION II: Change detection Publication D: Geomorphology


Page 98 PhD thesis

6.3. RESULTS

6.3.1. Geomorphological evolution of the
rock face
The five main events that occurred during the
period of study (A to E) are shown in Fig. 6.3. This
figure also shows the areas of deposition of events
A and D. The greater the dimensions and thickness
of the area with D
i
positive values, the greater the
volume of the rockfall(s).

Event A has the largest magnitude (87 m
3
).
Volumes of events B to E range from 1 to 10 m
3
.
Events A, B, D and E were clearly controlled by
open discontinuities, sub-parallel to the rock face
and visible in the field. As many researchers have
stated (e.g. Bieniawski, 1973; Romana, 1988), open
discontinuities are a clear conditioning factor of
failure. Furthermore, open discontinuities may
reveal a progressive movement of certain parts of
the rock face towards the external part of the slope.

The period during which the different events (from
A to E) took place can be seen in Fig. 6.4a. This
figure also shows the detected pre-failure
deformation for event B and an ongoing pre-failure
deformation in another part of the rock face: area F.
Fig. 6.4b shows the daily cumulated rainfall at the
nearest meteorological station (Sant Romà d'Abella).

Table 6.1 shows the cumulated rainfall for a given
interval. The “mean rainfall”, defined as the ratio
between the cumulated rainfall and a given period
of time can be seen in this table. This table also
shows the number of rockfalls for a given period of
time. The “rockfall activity” is defined as the ratio
between the number of new events and a given
period of time. A linear relationship between
rockfall activity and mean rainfall can be seen in
Fig. 6.4b. On the one hand, a dry season (low
cumulated daily rainfall) and low rockfall activity is
observed on the second row of table 6.1 (2nd
interval). On the other hand, the higher the mean
rainfall, the greater the rockfall activity.

A visual report of all the events that took place
during the study can be seen in Fig. 6.5. The
comparison of the sequential datasets during the
periods i, ii and iii can be seen in Fig 6.5a, b and c,
respectively. Period i is characterized by event A
and a few small scale rockfalls (less than 0.02m
3

each) on the left of the figure, showing minor
rockfall activity in period ii including event B.
Period iii is characterized by events C, D and F with
considerable rockfall activity.


6.3.2. Spatial prediction of rockfall
The spatial prediction of rockfall is based on the
detection of the pre-failure deformation. In order to
spatially detect precursory displacements over the
rock face, a detailed mesh (one point each 5 cm)
was generated in different sections of the rock face.
As stated in Section 2.4, the scattering of the
instrumental values was reduced by applying the
NN filtering technique.

Two areas showed a centimetric deformation: areas
B and F (Figs. 6.6 and 6.7 respectively). This
deformation was calculated along the Y axis
direction and was identified as a slope creep process.
Both of these displaced areas showed the following
characteristics: (a) A sub-vertical fracture
delimiting the moving part from the rest of the slope.
(b) An increase in the horizontal displacement
upwards (see Fig. 6.6 and 6.7). This type of
displacement is characteristic of a toppling
deformation. (c) A maximum value of the pre-
failure deformation of ~0.06 m. The pre-failure
deformation in area B was detected in period i (Fig
6.6b). This deformation was followed by a failure in
this area (Event B). Based on Terzaghi’s (1950)
observations for landslides, a causal relationship
between the pre-failure deformation and the minor
rockfall event was observed.

As regards area F, the pre-failure deformation is not
yet visible in period i (Fig. 6.7b). It started to appear
in period ii (Fig. 6.7c). As soon at its value doubles
the standard deviation, this deformation is clearly
visible in period iii (Fig. 6.7d). The maximum value
of the deformation is similar to that of event B with
the result that a failure is very likely.



6.4. DISCUSSION

The first part of the research deals with the
geomorphological evolution of the rockslope during
the study and evaluates the dimensions and
geometry of the main events and rockfall activity.
After the 1881 landslide event, the end of horizontal
stress generated tensional cracks that expanded
horizontally. These discontinuities constitute the
surface of detachment for the subsequent minor
failures. Despite the occurrence of rockfall and
rockslides, toppling is the main type of failure in the
study area.






Abellán et al. (accepted) Chapter 6: Rockfall detection and prediction




Antonio Abellán Fernández Page 99



Fig. 6.4. (a) Bar diagram showing the intervals where the different events occurred (from A to F?). The
intervals with the precursory displacements of the predicted events B and F? are also provided. (b) Daily
rainfall during the period of study (Sant Romà d'Abella, from SMC). The relationship between daily
events and mean daily rainfall is depicted by a star symbol (see text for further explanation).





Days
New
events
Rockfall frequency
(events /days)
Cumulated
rainfall (mm)
Mean rainfall
(mm / days)
1
st
interval 59 7 0.119 123.6 2.095
2
nd
interval 76 6 0.079 36.6 0.482
3
rd
interval 163 29 0.178 579.0 3.552

Table 6.1 Rainfall vs. rockfall frequency





SECTION II: Change detection Publication D: Geomorphology


Page 100 PhD thesis





Fig. 6.5. Comparison of sequential scans showing the evolution of the rock face. (a) period i; (b) period ii; and (c) period iii. Each of
the new rockfalls is indicated by a white circle.





Abellán et al. (accepted) Chapter 6: Rockfall detection and prediction




Antonio Abellán Fernández Page 101

Mechanically, the rotation of a block on a fixed
base (Muller, 1968) progressively reduces the safety
factor. Once the centre of gravity of the block
moves beyond the edge of rotation, the system is
destabilized, giving rise to an imminent toppling
failure. Not surprisingly, an influence of the
variable rainfall on the rockfall activity was
observed: the higher the cumulated rainfall in a
given interval, the greater the rockfall activity.

The second part of the study is focused on the
spatial prediction of rockfalls through the detection
of the slope creep. Almost 60 years ago, Terzaghi
(1950) observed the surface movements preceding a
landslide and stated: “It has often stated that certain
slides occurred without warning […] If a landslide
comes as a surprise to the eyewitnesses, it would be
more accurate to say that the observers failed to
detect the phenomena which preceded the slide…”.
This pre-failure deformation has been extensively
studied for soil and rock landslides in several
million m
3
(e.g. Saito, 1969; Voight, 1989 Crosta
and Agliardi, 2004) . However, apart from the
works of Zvelebill and Moser (2001) involving
large scale rockfalls (1.3·to 6×10
4
m
3
) and Rosser et
al. (2007) on the detection of decimetric rockfalls
leading to larger scale events, there is scant
literature on the prediction of minor scale rockfalls,
such as those discussed in this study. The amount
and duration of the pre-failure deformation have a
positive relationship with the dimensions of the
slide in ductile materials (Maruyama and Kozima,
1994). The visco-elastic behaviour of soils allows
the absorption of part of the stress and transforms it
into the deformation of the terrain. By contrast, the




Fig. 6.6. Centimetric displacement as a precursory indicator to event B. (a) picture of the studied area showing the area of
comparison. (b) scan comparison in period (i) using the 5x5 filtering technique. A centimetric displacement is observed at the same
area where event B took place in period (ii). The implications of this precursory displacement can be found in the text.






SECTION II: Change detection Publication D: Geomorphology


Page 102 PhD thesis

brittle behaviour of rocks does not allow much
deformation before the rupture of the material.
These circumstances may explain the scarce
detection of pre-failure deformation of minor
rockfall. One question that remains to be resolved is:
are all rockfalls preceded by a precursory indicator?
Drawing an analogy with the observations of
Terzaghi (1950) on soil slides, we can also say: are
we failing to detect the phenomena that precede
minor rockfalls?

In our case study, a causal relationship was
established in area B: a precursory deformation of a
few cm leading to the event B rockfall. Moreover, a
growing pre-failure deformation was detected in
area F in periods i and ii. Based on the assumption
of Terzaghi (1950), there is a high probability of
rockfall in the near future (event F?).

Finally, no precursory indicator was detected prior
to events A, C, D and E (Fig. 6.8). This could be
due to a combination of the following reasons:

(a) The periods analysed were too long: no
detection is possible if temporal resolution of
the study is shorter than the slope creep length.
Cliff evolution was not measured continuously.
For this reason, the exact date of each rockfall
event is unknown in our study. A more
frequent acquisition of datasets would have
enabled us to register more examples of
precursory deformation leading to a rockfall.
This could be the most important reason for the
lack of pre-failure deformation in C to E events.
(b) The accuracy of the method of study could be
lower than the real displacement. This may
also apply to C to E events.
(c) A lack of information prior to this research:
event A occurred somewhere between the 1st
and 2nd data acquisitions (period i). Because
no prior surface of reference was acquired,
slope creep detection was impossible.



Fig. 6.7. Progressive pre-failure deformation in area F. (a) study area and analyzed section. (from b to d) comparison of datasets: period i, ii
and iii, respectively. Colour scale from -0.1 up to 0.1m.







Abellán et al. (accepted) Chapter 6: Rockfall detection and prediction




Antonio Abellán Fernández Page 103

(d) Finally, in the worst scenario, an absence of
precursory indicators could also be a
possibility. However, as stated above, the pre-
failure deformation pattern has been
extensively observed in large scale soil slides
and rockfalls. Furthermore, this pattern was
observed in a previous event in the same study
area (event B).



6.5. CONCLUSIONS

A total of 42 rockfalls ranging from 10
-3
to 87 m
3

were recorded in the 10-month study period.
Dry/humid intervals correspond to lower/higher
rockfall activity, respectively. The detection of a
pre-failure deformation in area B enabled the
prediction of a toppling event (event B). Moreover,
an ongoing displacement of a few centimetres was
observed in area F. Its value increased in the upper
part of the moving block meaning that a toppling
failure is very likely (event F?). These results
indicate that TLS can be used for spatial prediction
of minor rockfalls over large areas.


ACKNOWLEDGEMENTS
Financial support from the Spanish Ministry of
Science and Education (pre-doctoral grant AP-
2007-1852) is gratefully acknowledged. This work
was also funded by Geomodels Institute, the project
TopoIberia CSD 2006 - 0004 / Consolider - Ingenio
2010 and the MEC project CGL2006-06596
(DALMASA). The authors wish to acknowledge
David Garcia for his support in the TLS data
acquisition. George von Knorring corrected the
English of the manuscript. Daily rainfall was
provided by the SMC (Servei Metereològic de
Catalunya).


REFERENCES

Abellán, A., Vilaplana, J.M., Martínez, J., 2006. Application
of a long-range Terrestrial Laser Scanner to a detailed
rockfall study at Vall de Núria (Eastern Pyrenees, Spain).
Engineering Geology 88, 136–148.
Abellán, A., Jaboyedoff, M., Oppikofer, T., Vilaplana, J. M.,
2009. Detection of millimetric deformation using a
terrestrial laser scanner: experiment and application to a
rockfall event, Nat. Hazards Earth Syst. Sci., 9, 365-372.
Agliardi, F., Crosta, G.B., 2003. High resolution three-
dimensional numerical modelling of rockfalls.
International Journal of Rock Mechanics and Mining
Sciences 40, 455–471.
Bieniawski, Z.T., 1973. Engineering classification of jointed
rock masses. Trans. S. African Inst. Civil Engineers 15,
335–342.
Blanchard, J., Calvet, J., Abellán, A., García, D., Khazaradze,
G., Vilaplana, J.M., 2008. Estudio del escarpe del
deslizamiento de Puigcercós mediante láser escáner

Fig. 6.8 Bar diagram showing an interpretation of the results: measured vs. interpreted pre-failure deformation (further explanation is
provided in the discussion section).







SECTION II: Change detection Publication D: Geomorphology


Page 104 PhD thesis

terrestre. Conca de Tremp, Catalunya. Geotemas 10,
1389-1392.
Chen, Y., Medioni, G., 1992. Object Modelling by
Registration of Multiple Range Images. Image and Vision
Computing 10, 145–155.
Copons, R., Vilaplana, J.M., Corominas, J., Altimir, J.,
Amigó, J., 2005. Rockfall risk management in high-
density urban areas. The Andorran experience. In: Glade,
T., Anderson, M., Crozier, M. (Eds.). Landslide Hazard
and Risk. John Wiley & Sons Ltd, Chichester, pp. 675-
698.
Copons, R., Vilaplana, J.M., 2008. Rockfall susceptibility
zoning at a large scale: From geomorphological inventory
to preliminary land use planning. Engineering Geology
102, 142-151.
Corominas, J., Alonso, E., 1984. Inestabilidad de laderas en el
Pirineo catalán. Tipologia y causas. Inestabilidad de
laderas en el Pirineo. Ponencias y Comunicaciones.
ETSICCP-UPC, C.1-C.53.
Crosta, G.B., Agliardi, F., 2003. Failure forecast for large
rock slides by surface displacement measurements.
Canadian Geotechnical Journal 40, 176-191
Dunning, S.A., Massey, C.R., Rosser, N.J., 2009. Structural
and geomorphological controls on landslides in the
Bhutan Himalaya using Terrestrial Laser Scanning.
Geomorphology 103, 17-29.
Frattini, P., Crosta, G., Carrara, A., Agliardi, F., 2008.
Assessment of rockfall susceptibility by integrating
statistical and physically-based approaches.
Geomorphology 94, 419-437.
Fell, R., Corominas, J., Bonnard, C., Cascini, L., Leroi, E.,
Savage, W.Z., 2008. Guidelines for landslide
susceptibility, hazard and risk zoning for land use
planning. Engineering Geology 102, 85-89.
Goodman, R.E., Bray, J.W., 1976. Toppling of rock slopes. In:
Rock Engineering for Foundations and Slopes;
Proceedings of a Specialty Conference, Vol. 2, Am. Soc.
Civ. Eng., New York, Pp. 201-233.
Guzzetti, F., 2000. Landslide fatalities and the evaluation of
landslide risk in Italy. Engineering Geology 58, 89–107.
Guzzetti, F., Reichenbach, P., Ghigi, S., 2004. Rockfall
hazard and risk assessment along a transportation corridor
in the Nera Valley, Central Italy. Environ. Manage. 34,
191–208.
Jaboyedoff, M., Metzger, R., Oppikofer T., Couture, R.,
Derron, M.-H., Locat, J., Turmel, D., 2007. New insight
techniques to analyze rock-slope relief using DEM and
3D-imaging cloud points: COLTOP-3D software. In:
Eberhardt, E., Stead, D., Morrison, T. (Eds.) Rock
Mechanics: Meeting Society's Challenges and Demands
(Vol. 2). Taylor & Francis, pp. 61-68.
Lichti, D.D., 2007. Error modelling, calibration and analysis
of an AM-CW terrestrial laser scanner system, ISPRS
Journal of Photogrammetry and Remote Sensing 61 (5),
307–324
Lim, M., Petley, D.N., Rosser, N.J., Allison, R.J., Long, A.J.,
Pybus, D., 2006. Combined digital photogrammetry and
time-of-flight laser scanning for monitoring cliff
evolution. Photogrammetric Record 20(1), 109-129.
Malamud, B.D., Turcotte, D.L., Guzzetti, F., Reichenbach, P.,
2004. Landslide inventories and their statistical properties.
Earth Surface Processes and Landforms 29, 687–711.
Muller, L., 1968. New considerations on the Vajont slide.
Felsmechanik und Ingenieurgeologie 6 (1), 1–91.
Maruyama, K., Kozima, S., 1994. A proposal of the warning
criterion for the landslide crisis depended on the
observation data of sliding distance. Journal of Japan
Landslide Society 31(1), 45-51.
Oppikofer, T., Jaboyedoff, M., Keusen, H.-R., 2008. Collapse
of the eastern Eiger flank in the Swiss Alps. Nature
Geosciences 1, 531-535.
Petrie and Toth, 2008. Introduction to laser ranging, profiling
and scanning. In Topographic Laser Ranging and
Scanning: principles and processing. Edited by: Shan, J.,
Toth, C. K., CRC Press, Taylor & Francis, 590p.
Romana, M., 1988. Practice of SMR classification for slope
appraisal. In: Bonnard, C. (Ed.), Proc. 5th International
Symposium on Landslides, Lausanne. Balkema,
Rotterdam. Vol 2, 1227-1232.
Rose, N.D., Hungr, O., 2007. Forecasting potential rock slope
failure in open pit mines using the inverse-velocity
method. International Journal of Rock Mechanics and
Mining Sciences 44, 308-320.
Rosser, N.J., Petley, D.N., Lim, M., Dunning, S.A., Allison,
R.J., 2005. Terrestrial laser scanning for monitoring the
process of hard rock coastal cliff erosion. Quarterly
Journal of Engineering Geology and Hydrogeology 38,
363-375.
Rosser, N.J., Lim, N, Petley, D.N., Dunning, S., Allison, R.J.,
2007. Patterns of precursory rockfall prior to slope failure.
Journal of Geophysical Research 112, No. F4.
Saito, M., 1969. Forecasting time of slope failure by tertiary
creep. Proceedings 7th International Conference on Soil
Mechanics and Found Engineering 2, 677−683.
Stoffel, M., Schneuwly, D., Bollschweiler, M., Lièvre, I.,
Delaloye, R., Myint, M., Monbaron, M., 2005. Analyzing
rockfall activity (1600–2002) in a protection forest – a
case study using dendrogeomorphology. Geomorphology
68, 224–241.
Szwedzicki, T., 2003. Rock mass behaviour prior to failure.
International Journal of Rock Mechanics & Mining
Sciences 40, 573-584
Terzaghi, K., 1950. Mechanism of landslides. In: Paige, S.
(Ed.) Application of Geology to Engineering Practice
(Berkeley Volume). Geological Society of America,
Washington D.C., pp. 83–123.
Teza, G., Galgaro, A., Zaltron, N., Genevois, R., 2007.
Terrestrial laser scanner to detect landslide displacement
fields: a new approach. International Journal of Remote
Sensing 28, 3425–3446.
Van Westen, C.J., Van Asch, T.W.J., Soeters, R., 2006.
Landslide hazard and risk zonation—why is it still so
difficult? Bull Engineering Geology and Environment 65,
167–184.
Varnes, D.J., 1978. Slope movement types and processes. In:
Schuster, R.L., Krizek, R.J. (Eds.), Special Report 176:
Landslides: Analysis and Control. TRB, National
Research Council, Washington D.C., pp. 11–33.
Vidal, L.M., 1881. Nota acerca de los hundimientos ocurridos
en la Cuenca de Tremp (Lérida) en Enero de 1881.
Boletin de la Comisión del Mapa Geológico de España,
Tomo VIII, 113-129.
Voight, B., 1989. A relation to describe rate-dependent
material failure. Science 243, 200–203.
Wyllie, D.C., Mah, C.W., Hoek, E., 2004. Rock Slope
Engineering: Civil and Mining. Taylor & Francis, 431 p.
Zvelebill, J., Moser, M., 2001. Monitoring Based Time-
prediction of rock falls: Three Case-Histories. Phys.
Chem. Earth (B) 26, 159–167.





III
SECTION

Global summary











































































CHAPTER 7:
Summary of
the results







Chapter 7. Summary of the results


Antonio Abellán Fernández Page 109














A brief summary of the results of this research can be found in the following table (7.1).

PUBLICATION
(Chapter)
DETECTION OF
PRECURSORY
DEFORMATION
ROCKFALL CHARACTERIZATION
(detachment)
EVOLUTION OF THE
PHENOMENON
Pub. A
(Chapter 2)
-
Study of the discontinuity planes,
volumes and rockfall source areas
during the period of study (back
analysis procedure).

Improvement of the rockfall
simulations using a high
accuracy and high
resolution Digital Elevation
Model (DEM).

Pub. B
(Chapter 4)
-
Calculation of rockfall volumes and
frequencies occurring in the
monitoring period. Estimation of the
rate of retreat.
-
Pub. C
(Chapter 5)
Development of methodologies
for the detection of precursory
deformation. Study of
precursory deformation in a
real case study (back analysis
procedure).
- -
Pub. D
(Chapter 6)
Rockfall detection and
prediction.
Calculation of rockfall volumes and
frequencies occurring in the
monitoring period.
-
Table 7.1 Summary of the results in each Publication (Chapters 2, 4, 5 and 6). These results concern the different
rockfall stages: precursory deformation, detachment and evolution of the phenomenon.



SECTION III. Change detection



Page 110 PhD thesis
7.1 Main applications of TLS in rockfall studies.
On the one hand, a single fieldwork campaign (Chapter 2) has allowed us to identify some of the
main parameters that control rockfalls. On the other hand, multiple data acquisition has allowed us
to study the geomorphological changes of a slope in a given area, during a certain time period.


7.1.1 Data acquisition in a single fieldwork campaign
The acquisition of TLS datasets in a single fieldwork campaign (Chapter 2) allowed us to obtain the
following parameters:

(a) the coordinates of the detachment area (section 2.4.1.1), estimated using traditional methods
(visual recognition) as well as TLS datasets, the latter being more accurate than the former;
(b) the orientation (dip direction and dip angle, section 2.4.1.2) of the joints that characterize the
bedrock, calculated using the methodology described in Fernández (2004). In contrast to
conventional methods (compass), direct contact between operator and rock face was
avoided. If the facets of a rock outcrop are determined by discontinuities, semi-automatic
methods could be carried out (Sturzenegger and Stead, 2009; Lato et al., 2009). However,
in the Vall de Núria case study, user selection along the points of the same joint yielded
better results;
(c) the volume of the detached blocks, estimated using the distance, positions and angles
between different joints;
(d) a more accurate estimate of rockfall trajectories and energies, obtained using a high-
resolution and high-accuracy Digital Elevation Model (DEM). This result is due to the fact
that micro-topographic and 3D variations of the concavities and convexities of the slope
played an important role in the dynamics of rockfall paths (e.g. Giani et al., 2004; Crosta
and Agliardi, 2004), making the exact determination of the trajectories unlikely. These
results were obtained employing a commercial software called Rotomap (Scioldo, 1991);



7.1.2 Multi-temporal data acquisition
Multi-temporal TLS data acquisition allows the geomorphological evolution of a given area to be
studied, such as rockfalls (sections 4.3 and 6.3.1) and small displacements (sections 5.3 and 5.4).
The latter have been identified as precursory rockfall indicators. Hence, these indicators have been
used to predict the location of future rockfalls (sections 6.3.2 and 7.4). Since the emphasis of this
thesis is on detecting change using TLS, the results from multiple data acquisition will be
extensively discussed in the following three sections.





Chapter 7. Summary of the results


Antonio Abellán Fernández Page 111
7.2 Rockfall detection
Sequential TLS datasets allow pre- and post-failure surfaces to be compared. As a result, the
morphology, magnitude, frequency and location of the detachment area can be obtained. To this end,
two different pilot study areas were monitored: Castellfollit de la Roca (Chapter 4) and Puigcercós
(Chapter 6).


7.2.1. Castellfollit de la Roca study area
This site was monitored for 22 months (from March 2006 to January 2008). The following results
were obtained: (a) Detachment of single columns: The detachment of 6 single columns was the
mass movement of highest frequency and lowest magnitude in the study area. The magnitude of each
of these rockfalls was below 1.5 m
3
, the geometry being controlled by the columnar jointing pattern.
The detachment of a single column can be seen in Figure 7.1a. (b) Detachment of a group of
columns: In April 2007 (event i; Fig. 7.1c; magnitude, 50 m
3
), a failure that affected a group of
columns in the central part of the rock face (Fig. 7.1b) was recorded. As a result of this rockfall, the
upper part of the cliff lost part of its basal support. Hence, a future event in this area is likely. The
average cliff retreat rate (e.g. Krautblatter and Dikau, 2007; Rabatel et al., 2008; Lim et al., 2009)
was estimated for the Castellfollit de la Roca study area using a combination of historical records
and results from the TLS monitoring campaign (see section 4.4). Its value ranged from 6 to 11
mm/year, assuming a detachment of material from 46 to 91.5 m
3
/year (see table 4.4) and a rock face
area of ~8.000 m
2
.

Fig. 7.1. Comparison of the sequential TLS datasets. Each point of the 3D point cloud is colour-coded in
accordance with the changes recorded during the period of comparison (a) Pre- and post-failure stages of the
detachment of a single column are visible; (b) event i; (c) event ii.



SECTION III. Change detection



Page 112 PhD thesis
7.2.2 Puigcercós study area
This site was monitored for 10 months (from September 2007 to July 2008). Data acquisition was
repeated 59, 135 and 303 days after the initial acquisition (Periods: P
i
, P
ii
, P
iii
, respectively). The
Puigcercós study area was characterized by higher rockfall activity (more than 40 rockfalls from 0.1
to 5 m
3
) than Castellfollit de la Roca (see previous section). Furthermore, the highest magnitude
event recorded during this research occurred in Puigcercós (volume = 87 m
3
). As an example of the
change detection method for rockfall detection, Figure 7.2 shows the detachment and deposition
areas of the main events that occurred during the time period. Figure 7.3 shows the comparison of
the sequential TLS datasets obtained during the monitoring period at this area. The events that
occurred in the last study period (Piii) are indicated by a white circle. Periods Pi and Pii are
characterized by low rockfall frequency. Piii is characterized by considerable rockfall frequency and
the events C, D and E (volumes ranging from 1 to 10 m
3
each, see Fig 7.2). A direct relationship




Fig. 7.2. Location and delimitation of the main rockfalls (from A to D) that occurred in Puigcercós during the
TLS monitoring campaign. See Chapter 6 for further information.



Chapter 7. Summary of the results


Antonio Abellán Fernández Page 113
between rockfall frequency and mean rainfall was obtained: the higher the mean rainfall, the greater
the rockfall activity. A graphical comparison of the periods Pi and Pii can be seen in Figure 6.5. A
more detailed analysis of these results can be found in section 6.3 (Publication D).



7.3 Detection of small displacements
This section summarizes the results of Chapter 5 (Publication C). The aim was to ascertain whether
TLS instrumental error is small enough to enable detection of precursory displacements of
centimetric and/or subcentimetric magnitude. To this end, an outdoor experiment was performed at
the University of Lausanne in collaboration with the Institute of Geomatics and Risk Analysis
(IGAR) in October 2007. The methods developed were applied to a real rockfall event.


7.3.1 Outdoor experiment
The experiment consisted of the simulated displacement of three objects: (i) a plane, (ii) a
hemisphere and (iii) an irregular form, relative to a fixed, stable and vertical plane (see Fig. 5.1). The
displacement of the three objects ranged between 5 and 25 mm, with an increment of 5 mm between
each scan. After each induced displacement, a TLS point cloud (referred hereinafter RAW data) was
acquired and compared with the initial point cloud captured at 0 mm displacement. In order to
reduce the error in RAW data comparison, a Nearest Neighbour (NN) averaging technique was also
applied.


Fig.7.3. Comparison of sequential scans showing the evolution of the rock face. Each new rockfall during period
Piii is indicated by a white circle.



SECTION III. Change detection



Page 114 PhD thesis

Fig. 7.4. (from a to c) RAW data displacements for different induced displacements (5, 10, 15 mm,
respectively); (from d to f) displacement measurements after the 5×5 Nearest Neighbour averaging for similar
displacements (5, 10 and 15 mm, respectively). The induced displacements are more accurately computed after
NN averaging.

(a) RAW data displacements: The variance of the RAW data values
(7.4 mm < σ < 11.2 mm) were of the same order of magnitude as the first and second
deformation increments of the experiment (5 and 10 mm, respectively). As a result, the real
displacements in Figures 7.4a and 7.4b (5 and 10 mm, respectively) were masked by the
scattering of the RAW data. In these figures, it is difficult to differentiate moving and fixed



Chapter 7. Summary of the results


Antonio Abellán Fernández Page 115




Fig. 7.5. Filtered cumulated displacement using the NN averaging technique for a 6-month comparison prior to
the 50 m
3
rockfall. A centimetric precursory displacement is observed in the middle of the figure.
parts. The induced displacements can be detected at 15 mm (Fig. 7.4c). Finally,
displacements were clearly visible at 20 and 25 mm.

(b) Nearest Neighbour averaging: Instrumental error was filtered out using this method,
allowing the real displacement to become visible in Figures 7.4d to 7.4f (5, 10 and 15 mm.,
respectively). In contrast to the low precision (or high variability) obtained using
conventional methods (Fig. 7.4a to 7.4c), displacements were more accurately computed
after NN averaging, even for the smallest displacement value (5 mm). The standard
deviation of the comparison model was up to six times more accurate using NN averaging
than with conventional methods (σ
NN
= 1.3 cm vs. σ
RAW
= 7.2 cm). For this reason, the
technique was applied to a real case study (see next section).


7.3.2 Application to a rockfall event
A set of 50 m
3
columnar basalt blocks fell from the Castellfollit de la Roca rock face in April 2007.
A comparison of the reference and data point clouds using the RAW data suggested that the cliff
underwent deformation in the 6 months prior to the failure. However, owing to the high variability
discussed above, we were unable to clearly delimit the extent of deformation. The NN averaging
technique was applied to the same TLS datasets, allowing a maximum precursory displacement of
45 mm (Fig. 7.5) to be detected. The detection of the precursory deformation in this real case study



SECTION III. Change detection



Page 116 PhD thesis
was performed through a back analysis procedure. The location of the likely rockfall was not
predicted. Nevertheless, this case study opens up the possibility for the spatial prediction of rockfalls
through the early detection of small-scale displacements.



7.4 Spatial prediction of rockfalls
Precursory deformation in small-scale rockfalls was detected in the Castellfollit de la Roca and
Puigcercós study areas. Of these deformations, (a) growing deformation across two areas culminated
in a rockfall occurrence (areas i and B, figures 5.5 and 6.6 respectively); and (b) another growing
deformation across three areas (areas F, G, H, Fig. 7.6) was detected, making a subsequent rockfall
likely.


























The main characteristics of these areas are provided below (see table 7.2).


Fig. 7.6. Location of the areas with precursory deformation (Areas F, G, H). The moving areas are delimited by a sub-vertical
fracture (white line). Area F is enlarged and rotated.





Chapter 7. Summary of the results


Antonio Abellán Fernández Page 117

Precursory deformation Name
(Location)
Volume
(m
3
) Max. extension
(cm)
Period
(days)
Detachment of
the material?
(Yes/No)
Early prediction
vs. back analysis
Event i
(Castellfollit de la Roca)
50 4.5 P
3 a 4
(182) YES Back analysis
Area B
(Puigcercós)
2 6 P
i
(57) YES Early prediction
Area F
(Puigcercós)
- 6 P
i a v
(546) - Early prediction
Area G
(Puigcercós)
- 6 P
i a v
(546) - Early prediction
Area H
(Puigcercós)
- 4 P
i a v
(546) - Early prediction
Table 7.2. Areas with precursory deformation detected in 2007-2008.


The precursory deformation detected in Castellfollit de la Roca and Puigcercós showed an increase
in the horizontal displacement with time. The areas with precursory deformations detected in
Puigcercós showed the following characteristics: (a) a sub-vertical fracture delimiting the moving
part from the rest of the slope; (b) an increase in the horizontal displacement upwards, typical of a
toppling failure mechanism (Muller 1968; Goodman and Bray, 1976). In addition, decimetric-scale
rockfalls were observed in the upper part of the moving areas. These results are consistent with the
observations of Rosser et al., (2007).

























SECTION III. Change detection



Page 118 PhD thesis






























CHAPTER 8:
Summary of
the discussion






SECTION III. General Summary



Page 120 PhD thesis







Chapter 8. Summary of the discussion



Antonio Abellán Fernández Page 121













This chapter is divided into four sections, as outlined as follows: (a) a critical analysis of the research
(strengths and weaknesses of TLS, comparison with related works and with related techniques) is
provided in Section 8.1; (b) a discussion of rockfall detection in different study areas (Castellfollit de
la Roca and Puigcercós) is presented in Section 8.2; (c) a discussion of small displacement detection
and its application to spatial prediction of rockfalls is provided in Section 8.3; (d) finally, Section 8.4
is a discussion of the spatial prediction of rockfalls through the detection of precursory
displacements, with an emphasis on the Puigcercós study area.



8.1 Critical analysis

8.1.1 Strengths and weaknesses of TLS

(a) The main strengths of TLS are as follows:

(i) high accuracy: 7.2 mm at a range of 50 meters (see table 5.1);

(ii) high angular resolution (e.g. 1 point every 5 mm, see Section 5.2.2);

(iii) fast data acquisition: 2,500 points/second (Optech, 2009);

(iv) broad coverage: characterization of large vertical surfaces (e.g. 8·10
3
m
2
in Castellfollit,
see Chapter 4), meaning a large quantity of information can be obtained (several million
points in each TLS dataset, see Chapters 4, 5 and 6);

(v) considerable maximum range on natural slopes (usually 600-700 m; maximum range in
our study: 870 m, see Chapter 2);
(vi) intermediary prism reflectors are not necessary, allowing inaccessible and dangerous
areas to be studied.



SECTION III. General Summary



Page 122 PhD thesis



(b) Some weaknesses were also encountered:

(i) the scanned surface may be misrepresented depending on the orientation of the TLS to the
rock face (Lato et al., 2009).

(ii) direct visibility (line-of-sight) is, as with other surveying, radar and/or optical techniques,
necessary between the TLS and the scanned surface: range measurements are erroneous in
the presence of vegetation, wires, etc;

(iii) the TLS ILRIS-3D wavelength does not allow 3D information to be acquired in the case
of snow or water seepage;

(iv) different TLS models allow datasets to be acquired at greater maximum ranges (e.g. up to
40% more using the ILRIS-3D Extended Range, Optech, 2009);

(v) the vector of comparison was defined in our study using a direction perpendicular to the
rock face, however, the direction of displacement may be different in a real case study
(e.g. vertical direction in the case of a collapse or falling). A real 3D field of displacement
could be an alternative method (e.g. Monserrat and Crosetto, 2008; Oppikofer et al.,
2009);

(vi) as regards the detection of subcentimetric deformation, instrumental error could be higher
than that of pre-failure displacement. As a result, the scattering of the TLS values could
mask the real pre-failure displacement (see Chapter 5);

(vii) undoubtedly, the most important limitation is the temporal variable: the short time-span
of the TLS campaign (specifically for the study of low rockfall frequencies) represented a
clear bias for the detection of large-scale failures in the Castellfollit de la Roca case study;

(viii) a more frequent dataset acquisition would have enabled us to register more examples of
precursory deformation leading to a rockfall in the Puigcercós case study (Section 6.3.2)
and better understand the influence of climatic variables as triggering factors (e.g. Lim et
al., 2009);



8.1.2 Comparison with related works

A comparison of this research with related works is as follows:

(a) Rockfall detection: In Chapter 4 and the first part of Chapter 6 (rockfall detection by TLS), the
methodologies developed by Rosser et al., (2005) and Lim et al., (2006) are applied to different
study areas (Castellfollit de la Roca and Puigcercós). Compared with these and related works
(e.g. Lim et al., 2009; Dewez et al., 2009) and including the research carried out at the
Puigcercós study area (see Chapter 6), our ability to estimate the rate of retreat at Castellfollit de
la Roca was limited by the scarce number of rockfalls (see Section 4.4.3). It goes without saying



Chapter 8. Summary of the discussion



Antonio Abellán Fernández Page 123



that a longer TLS monitoring period would increase the accuracy of this estimation (see “further
research” Section). Moreover, the continuous acquisition of TLS datasets over ongoing failures
would help to understand mass movement dynamics. As an example, monitoring the different
failure stages of the collapse that occurred on the eastern flank of the Eiger (Swiss Alps) allowed
a more accurate interpretation of the failure mechanism, depth of the stable area, etc. (Oppikofer
et al., 2008a).

(b) Rockfall precursory indicators: Related works dealing with the detection of precursory
deformation focus on temporal forecasting rather than spatial prediction (e.g. Saito, 1969;
Voight, 1989; Glawe et al., 1991; Lotter et al., 1998; Zvelebil and Moser, 2001; Crosta and
Agliardi, 2004; Rose and Hungr, 2007). Most of the previously mentioned studies all involved
back-analyses of slope movements; reported cases of forward prediction are few (Van Asch et
al., 2007). Except for the work of Rosser et al. (2007), which is based on an increase in rockfall
activity as a precursory indicator of larger failures, there is scant literature on the prediction of
minor-scale rockfalls.

(c) Spatial prediction of rockfalls: The concept of spatial prediction used in this work differs
significantly from that of other studies (e.g. Brabb et al., 1972; Carrara et al, 1983; Corominas,
1993; Lee et al., 2006). The latter is normally referred to as failure susceptibility or “probability
of occurrence” and does not include a time frame. This parameter is usually calculated by
processing the spatial information contained in a database (e.g. the orientation of discontinuities,
morphology of the slope, lithology, etc.) through diverse techniques (e.g. Bayesian probability,
neural networks, likelihood ratios, fuzzy sets, etc.), with the aim of detecting pixels with the
same characteristics of a slope movement that took place in the past. Although both approaches
can be complementary at different scales, our proposal offers:

(i) a much more precise delimitation of the moving area, obtained using a multi-temporal comparison of
TLS datasets;
(ii) a link between spatial and temporal prediction of rockfalls: once the precursory deformation is detected,
temporal forecasting of rockfalls may be performed using predictive models, such as the inverse velocity
method suggested by Saito (1969);

(d) Other applications: the development of new methods for the different applications of TLS
discussed in Chapter 2 is beyond the scope of this research, e.g. calculating the orientation of a
single discontinuity was based on Fernández (2004). Nowadays, these methods have been
extended to include the whole rock face (see Jaboyedoff et al., 2007; Pernito, 2008;
Sturzenegger and Stead, 2009; Lato et al., 2009). As stated elsewhere in this PhD dissertation,
after the first publication (Chapter 2), we have focused on the use of TLS for rockfall detection
and prediction.






SECTION III. General Summary



Page 124 PhD thesis



8.1.3 Comparison of TLS with other techniques
An overview of the accuracy and resolution of different monitoring techniques (e.g. geotechnical
instruments, surveying methods and remote sensing techniques) can be found in Settles et al. (2008).
Comparing TLS with some of these techniques, we find:

(a) Point-based monitoring techniques (e.g. tachometric measurements, real-time differential GPS,
extensometers, etc.): the main advantage of using TLS instead of these techniques is its high
resolution and capacity to effectively measure most of the rock face;

(b) Aerial LIDAR: other remote sensing methods of greater range, such as aerial LIDAR, cover
larger areas than TLS but provide less accuracy and density of information, e.g., vertical
accuracy and resolution of decimetric and metric order of magnitude (McKean and Roering,
2004). Moreover, with aerial LIDAR, it is not possible to acquire high density information in
areas with steep slopes, which are the areas characterized by higher rockfall susceptibility;
Nevertheless, a combination of terrestrial and aerial measurements (e.g. Ruiz et al., 2004;
Sturzenegger et al., 2007) could solve the occlusion problems, also called shadow areas (that is,
when the incidence angle is parallel to the orientation of the geological surface, Lichti et al.,
2007; Lato et al., 2009).

(c) Photogrammetry: TLS produces faster, more accurate and less expensive DEMs than terrestrial
photogrammetry (Bitelli et al., 2004);

(d) Ground-Based Synthetic Aperture Radar (GB-SAR): this technique may be an alternative
remote sensing technique for the detection of precursory displacements over large areas.
Different authors have satisfactorily employed a GB-SAR for the detection of subcentimetric
deformation in million cubic landslides (Antonello et al., 2004; Corsini et el., 2006; Noferini et
al., 2007). This technology provides more accurate measurements (for subcentimetric values in
natural slopes, see Noferini et al., 2007) than TLS even after NN averaging (for centimetric
values, see Chapter 6). Furthermore, the maximum range is considerably greater than that of
TLS (from 2.000 to 3.000 m vs. 800 m using TLS, see Teza et al., 2008). However, GB-SAR
has three main weaknesses:

(i) sudden changes such as rockfalls usually cause loss of coherence in the radar wavelength;
(ii) the required infrastructure for a GB-SAR set-up is more complex than for TLS: a permanent station is
necessary;
(iii) spatial resolution of GB-SAR usually ranges from 1 to 10 meters (e.g. Tarchi et al., 2005; Corsini et
al., 2006; Teza et al., 2008), making metric scale rockfall detection impossible.

As regards the aforementioned limitations:

(i) TLS does not suffer from signal coherence loss, meaning the detection of decimetric to million cubic
meters is possible (e.g. Rosser et al., 2005; Oppikofer et al., 2009);
(ii) the easier set-up of TLS is one of its main advantages, although accuracy is higher with a permanent
TLS station;



Chapter 8. Summary of the discussion



Antonio Abellán Fernández Page 125



(iii) TLS spatial resolution ranges from centimetric to subcentimetric values, i.e., two orders of magnitude
greater than that obtained using a GB-SAR, making the detection of rockfalls with a volume above
0.001 m3 possible (see Rosser et al., 2005).

The joint use of TLS and GB-SAR techniques is discussed in Teza et al. (2008). A combination
of both methods (i.e. the subcentimetric deformation detected by GB-SAR in conjunction with a
3D model of the slope obtained by TLS) may help understand the moving areas.




8.2 Rockfall detection

Two different approaches were employed to estimate rockfall frequency and magnitude in the
Castellfollit de la Roca study area: a long-term approach and the TLS monitoring campaign. Both
approaches suffered from systematic errors in the estimation of rockfall frequency. On the one hand,
the long-term approach suffered from a bias in the recording of phenomena: only rockfalls above a
certain volume are normally recorded and/or detected by witnesses (Hungr et al., 1999); on the other
hand, the results of the TLS campaign were temporally biased: the low frequency of medium- and
large-scale rockfalls demanded a longer study period. A combination of these approaches allowed
for a better understanding of the frequency-magnitude relationship in the study area.

The different lithologies, morphology of the slopes and discontinuities played an important role in
rockfall behaviour:

(a) Lithology: the different lithologies of the Castellfollit de la Roca (i.e. basalts) and Puigcercós
(i.e. marls and sandstones) rock faces controlled the mechanical behaviour of the failure. Basalt
and sandstone detachment is a brittle fracture controlled by pre-existing discontinuities. Marls
are characterized by ductile behaviour;

(b) Morphology: the protruding morphology of certain parts of the rock face controls the possibility
of collapse (Lim et al. 2009). This was the case with the highest failure volume recorded in
Castellfollit de la Roca (1500 m
3
, Sept 1976) and the highest failure volume recorded there
during the TLS monitoring campaign (50m
3
, April 2007);

(c) Discontinuities: tension fractures played an important role in the detachment of the higher
failures: after the 1881 landslide in the Puigcercós study area. The disappearance of a horizontal
stress generated tension cracks that expanded horizontally. These discontinuities constituted the
detachment surface of subsequent failures, such as the 87 m
3
failure (see Fig 6.3) and those areas
where precursory deformation was detected;




SECTION III. General Summary



Page 126 PhD thesis



Finally, despite the direct relationship between rockfall frequency and mean rainfall observed in the
Puigcercós study area (see Section 6.3.1), it was not possible to estimate a rainfall threshold. The
highest magnitude rockfall (87 m
3
) occurred during heavy rainfall (35 mm) in period Pi (Oct-Nov
2007). The same maximum daily rainfall was reached three more times during the monitoring period,
although no rockfall higher than 5 m
3
occurred. The same pattern was observed at the Castellfollit de
la Roca study area. As discussed in Lim et al. (2009), the correlation between climatic variables and
rockfall occurrence is affected by the size of the event: the higher the volume, the lower the
correlation factor. Others factors appear to control the occurrence of high magnitude rockfalls (Lim
et al., 2009). Hence, a preparatory factor, e.g., a precursory displacement, may be necessary for this
type of rockfall to occur.




8.3 Precursory deformation

One question that needed to be considered before our research was as follows: is the instrumental
error of a TLS system small enough to detect precursory deformation on rock slopes? Under normal
circumstances, instrumental error may mask precursory deformation. Nevertheless, the experimental
case study shown in Chapter 4 allowed the scattering of the TLS point cloud values to be better
understood, making the detection of subcentimetric deformation possible after the application of a
Nearest Neighbour (NN) averaging technique.

With the aim of obtaining good agreement between accuracy and resolution, different k values were
tested in the experimental case study discussed in Chapter 5. Algorithms that involved low numbers
of NN (k=8, 3×3 NN) retained significant noise. By contrast, algorithms that involved a larger
number of NN (k=35, 6×6 NN) masked local-scale displacements. A k value of 24 (5×5 NN) was
selected as an optimal compromise. In line with Ingensand et al. (2006), the object surface
orientation influenced the accuracy of measurements: the highest errors were concentrated on those
surfaces that were most oblique to the incidence angle, e.g., around the external part of the
hemisphere. Furthermore, our results show how the error value increases with the complexity of the
shape.

Applying the NN averaging technique to the 50 m
3
rockfall at Castellfollit de la Roca (see Section
5.4), underlined the utility of this method for precursory deformation detection in real cases studies.
Nevertheless, the accuracy of the method for this natural slope was considerably lower than that
obtained under “optimal conditions” (with σ
NN_NaturalSlope
= 6.4 mm against σ
NN_ExperimentalTest
= 1.3
mm). This may be explained by the greater distance, different reflectivity of the material,
morphology of the scanned surface and angle of incidence (Petrie and Toth, 2008; Voegtle et al.,
2008; Abellán et al., 2009; Lichti, 2007, respectively). The application of this technique at the



Chapter 8. Summary of the discussion



Antonio Abellán Fernández Page 127



Puigcercós scarp allowed four additional areas to be detected characterized by an increasing
precursory deformation (see summary of the results in the previous Chapter).

As stated in Publication D, the NN technique required validation in more areas to test its
applicability under real conditions. In order to fulfil this validation, we carried out new fieldwork
campaigns in February and March 2009. These new datasets confirmed the evolution of precursory
deformation in the ongoing displacement discussed in Chapter 7 (Areas F, G, H) over 8 months (see
Rodriguez et al., accepted). Furthermore, these datasets revealed the existence of three new areas
(Areas I, J, K, see Fig. 8.1 and table 8.1) with a precursory deformation that increased in value over
time (one of these areas with precursory deformation resulted in a failure).





















Precursory deformation
Name Location
Volume
(m
3
)
Max.
extension
(cm)
Period (days)
Detachment of the
material?
(Yes/No)
Early prediction
vs. back
analysis
Event I
Puigcercós 0.8 4 P
iv
(203) Yes back analysis
Area J
Puigcercós - 3 P
iv a v
(250) - Early prediction
Area K
Puigcercós - 3 P
iv a v
(250) - Early prediction

Table 8.1. New areas with precursory deformation detected during fieldwork carried out after the acceptance of
Publication D. These new areas confirmed the possibility of early precursory deformation detection.




Fig. 8.1. Location of the areas with precursory deformation (Areas F, G, H, I, J, K). The moving areas are delimited
by a sub-vertical fracture (white line).



SECTION III. General Summary



Page 128 PhD thesis



8.4 Spatial prediction of rockfalls

The causal relationship between precursory deformation of a few cm and a rockfall event was
established in the Castellfollit de la Roca and Puigcercós study areas. Furthermore, different ongoing
precursory deformations were detected in the Puigcercós rock face. Five areas are currently
characterized by an ongoing deformation (Areas F, G, H, J, K). The occurrence of a rockfall is
highly likely in these areas.

Precursory deformation has been extensively studied in several million cubic meters of soil and rock
landslides (e.g. Terzaghi, 1950; Saito, 1969; Voight, 1988; Fukuzono, 1990; Crosta and Agliardi,
2004) . However, apart from the long-term research (13 years) of an ongoing failure of 7·10
3
m
3
in
the Austrian Alps (see Glawe et al., 1991; Glawe and Lotter 1996; Lotter et al., 1998; Zvelebil and
Moser, 2001), there is scant literature on the prediction of minor-scale rockfalls, such as those
discussed in this study. This may be because most of the aforementioned studies on landslide
forecasting employed point-based instruments of measurement (e.g. extensometers, total stations,
etc.) to monitor the moving area. Since the moving area is often an unknown factor, a method to
detect the unstable areas (i.e. real spatial prediction) is still required. Another reason may be that, as
Maruyama and Kozima (1984) stated, the amount and duration of the precursory deformation in soil
slides have a positive relationship with the dimensions of the landslide. The visco-elastic behaviour
of soils allows the terrain to become deformed. However, the brittle nature of rocks does not allow
much deformation before the rupture of the material. These circumstances may explain why
precursory deformation is rarely detected before a minor rockfall. One question that remains to be
resolved is: are all rockfalls preceded by a precursory indicator? Or, in accordance with the
observations of Terzaghi (1950) on soil slides: are we failing to detect the phenomena that precede
minor rockfalls?


Precursory deformation was mainly detected prior to a toppling failure mechanism in ductile
materials and rockfalls of considerable volume:

(a) Failure mechanism: a toppling is defined as the rotation of a block on a fixed base (Muller,
1968). Hence, this movement may be characterized by greater precursory deformation than
that observed in other failure mechanisms such as a fall or slide. Deformation prior to the 50
m
3
failure in Castellfollit de la Roca was considerably lower than deformation prior to the
ongoing displacement observed in the Puigcercós rock face, both in length (5 cm vs. 25 cm,
respectively) and duration (180 vs. 576 days, respectively, unpublished data).

(b) Lithology: given that both events mobilized a similar volume, the different lithologies found in
the Castellfollit de la Roca and Puigcercós rock faces may have important implications in the
evolution (length and duration) of the precursory deformation. On the one hand, the brittle
behaviour of the basaltic and calcarenitic blocks does not allow as much deformation before
the rupture of the material. On the other hand, the ductile behaviour of marl allows part of the
stress to be absorbed and transformed into the deformation of the terrain, increasing the length



Chapter 8. Summary of the discussion



Antonio Abellán Fernández Page 129



and duration of the precursory deformation. Another explanation may be that rockfall
frequency at Puigcercós was much higher than at Castellfollit de la Roca. Hence, the
frequency in detecting precursory deformation should be higher in the former than the latter.

(c) Volume: finally, the magnitude of the rockfall may also play a key role in the detection of
precursory deformation. In an attempt to increase the spatial and temporal scale of our results,
the long-term monitoring of two rockfalls were incorporated: a 7·10
3
m
3
and a 4.6·10
6
m
3

toppling. Detailed information of these failures can be found in the literature (Vibert et al.,
1988; CETE, 1990; Glawe et al., 1991; Glawe and Lotter, 1996; Lotter et al., 1998; Zvelebil
and Moser, 2001; Rose and Hungr, 2007). A scale dependency in the length and duration of
precursory deformation was observed both in our results and in these case studies: the higher
the magnitude of the rockfall, the longer the length of precursory displacement.




It is also important to mention that no precursory indicators were detected prior to many of the
rockfalls that occurred in the study areas. This could be for several reasons. One, the periods
between data acquisitions were too long: no detection is possible if temporal resolution of the study
is shorter than the slope creep length. A more frequent acquisition of datasets would have enabled us
to register more examples of precursory deformation leading to a rockfall. Two, in some cases, the
accuracy of the study method could be lower than the actual displacement. And three, in the worst
scenario, an absence of precursory indicators is a possibility. However, as stated above, the
precursory deformation pattern has been extensively observed in large-scale soil slides and rockfalls.

The application of TLS for the spatial prediction of rockfalls should be validated through: (a) the
continuation of the TLS monitoring campaign at the 5 areas which currently show ongoing
deformation (Areas F, G, H, J, K); (b) the selection of new case studies at different
geomorphological sites with different lithologies; and (c) the selection of new case studies with
different failure mechanisms (e.g. fall, slide). These tasks are of paramount importance to understand
the pre-failure behaviour of rockfalls and to implement these findings in an early warning system.









SECTION III. General Summary



Page 130 PhD thesis







Chapter 9. Concluding remarks


Antonio Abellán Fernández Page 131

































CHAPTER 9:
Concluding
remarks



SECTION III. General summary



Page 132 PhD thesis










9.1 Summary of the conclusions


MAIN APPLICATIONS OF TLS

The use of TLS on rock slopes has considerably improved our understanding of rockfall phenomena.
Different applications presented in this and related works are undergoing rapid development, mainly
in the following aspects: generation of a high–resolution, high–accuracy Digital Elevation Model
(DEM) of steep slopes; 3D simulation of rockfall trajectories; characterization of 3D discontinuities;
monitoring of rock slopes; detection of rockfalls (geometry, spatial location, magnitude and
frequency); detection of precursory deformation and spatial prediction of rockfalls.


ROCKFALL DETECTION

The sequential acquisition of TLS datasets in the Castellfollit de la Roca and Puigcercós pilot study
areas, allowed rockfalls to be accurately located, and their magnitude and frequency to be accurately
estimated.

The detachment of six basaltic columns (each of less than 1.5 m
3
) and the detachment of a
group of columns (with a total volume of 50 m
3
) were recorded at Castellfollit de la Roca
during the 22-month monitoring period.

A considerably higher rockfall frequency was recorded in the Puigcercós study area: a total
of 42 rockfalls ranging from 10
-3
to 87 m
3
were recorded during the 9-month monitoring
period. Dry (wet) spells corresponded to periods characterized by lower (higher) rockfall
frequency.



Chapter 9. Concluding remarks


Antonio Abellán Fernández Page 133

DETECTION OF SMALL DISPLACEMENTS

The use of TLS for the detection of small-scale deformation proved useful in artificial scenery,
considerably improving the accuracy of comparisons by applying a Nearest Neighbour (NN)
averaging technique, which allowed subcentimetric deformation to be detected (σ
NN
: 1.3 mm. vs.
σ
RAW
: 7.2 mm; at a mean range of 50 m);

The use of TLS to detect centimetric deformations proved useful at two natural slopes
(Castellfollit de la Roca and Puigcercós) over eight different areas (one and seven
respectively).

These small scale deformations were identified as precursors of rockfalls (slope creep,
Terzaghi, 1950).


SPATIAL PREDICTION OF ROCKFALLS

Of the eight areas mentioned above: (a) three suffered a rockfall; (b) an increase in the
horizontal displacement was observed in the other five areas, meaning that a rockfall is very
likely. In addition, decimetric-scale rockfalls were observed in the upper part of the five
moving areas, which is consistent with the precursory indicators observed by Rosser et al.,
(2007).

However, precursory deformation was not detected prior to each rockfall that occurred
during the monitoring period (but this could simply be due to infrequent data acquisition or
insufficient instrument accuracy).

The parameters of the precursory deformation correlated with the failure mechanism,
lithology and volume of the rockfall: higher values of length and duration of the precursory
deformation were found in the toppling failure mechanism, ductile materials and rockfalls
that involved considerable volumes. These results are consistent with observations in the
literature regarding rockfalls of higher magnitude and lower frequency (e.g.: Zvelebil and
Moser, 2001; Crosta and Agliardi, 2004; Hungr et al., 2007).

The findings of this research are of paramount importance for the early detection of
rockfalls: the prediction of the location could be linked with temporal forecasting based on
predictive models developed by different authors (e.g.: Saito, 1969; Voight, 1989;
Fukuzono, 1985). This may constitute a step forward in rockfall risk management.









SECTION III. General summary



Page 134 PhD thesis

9.2 Further research

This research opens up many more questions than it answers. Future research should consider:

- The continuation of the study in the five areas characterized by an ongoing deformation at
Puigcercós scarp (areas F, G, H, J, K). As discussed in previous sections, the short time span
and the low frequency of dataset acquisition represented a clear restriction on the
understanding we were able to gain of rockfalls.

- The detection of precursory displacements under different conditions, i.e., for different
lithologies, in different pilot study areas, with variable ranges, and characterized by different
failure mechanisms.

- Predicting the timing and location of future rockfalls. This is the most pressing challenge in
the use of the TLS in rockfall studies. The combination of the TLS with different
monitoring techniques (GB-RADAR, strain gauges, micro seismic detectors, etc.) may shed
light on the different precursory rockfall indicators (pre-failure deformation, precursory
rockfalls, crack propagation, etc.).

























APPENDICES
Appendix 1: Overall references
Appendix 2: List of tables
Appendix 3: List of figures
Appendix 4: Specific terminology
Appendix 5: Extended summary in Spanish



APPENDIX 1


Page 136 PhD thesis










































Overall References


Antonio Abellán Fernández Page 137
































Appendix 1
Overall references





APPENDIX 1


Page 138 PhD thesis





























Overall References


Antonio Abellán Fernández Page 139











References


- Abellán, A., Jaboyedoff, M., Oppikofer, T., Vilaplana, J. M., 2009. Detection of millimetric deformation using
a terrestrial laser scanner: experiment and application to a rockfall event, Nat. Hazards Earth Syst. Sci., 9,
365-372.
- Abellán, A., Vilaplana, J.M. Calvet, J., Blanchard, J., accepted. Rock face monitoring by Terrestrial Laser
Scanning: change detection and spatial prediction of rockfall. Geomorphology.
- Abellán, A., Vilaplana, J.M., Calvet, J., García-Sellés, D., 2008. Seguimiento (monitoring) de laderas
rocosas mediante un Láser Escáner Terrestre. Zona de estudio piloto en Castellfollit de la Roca. Cataluña.
Geo-temas 10, 1385-1388.
- Abellán, A., Vilaplana, J.M., Calvet, J., García-Sellés, D., Asensio, E., submitted. Rockfall monitoring by
Terrestrial Laser Scanning. Case study of the basaltic rock face at Castellfollit de la Roca (Catalonia,
Spain). Nat. Hazards Earth Syst. Sci.
- Abellán, A., Vilaplana, J.M., Martínez, J., 2006. Application of a long-range Terrestrial Laser Scanner to a
detailed rockfall study at Vall de Núria (Eastern Pyrenees, Spain). Engineering Geology 88, 136–148.
- Agliardi, F., Crosta, G.B., 2003. High resolution three-dimensional numerical modelling of rockfalls.
International Journal of Rock Mechanics and Mining Sciences 40, 455–471.
- Alba, M., Longoni, L., Papini, M., Roncoroni, F., Scaioni, M., 2005. Feasibility and problems of TLS in
modelling rock faces for hazard mapping. In: proceedings of the ISPRS workshop laser scanning 2005,
Enschede, Netherlands. WG II/3, III/4, V/5. Available at <www.commission3.isprs.org/
laserscanning2005>.
- Antonello, G., Casagli, N., Farina, P., Leva, D., Nico, G., Sieber, A., Tarchi, D., 2004. Ground-based SAR
interferometry for monitoring mass movements. Landslides 1(1):21–28.
- Araiba, K., 2006. Study on the Method for Detecting and Monitoring of Pre-Failure Deformation in Slope.
Proceedings of the INTERPRAEVENT International Symposium Disaster Mitigation of Debris Flows,
Slope Failures and Landslides held on September 25–27, 2006 in Niigata, Japan. Vol.2, 581-589.



APPENDIX 1


Page 140 PhD thesis
- Armesto, J., Ordóñez, C., Alejano, L., Arias, P., 2009. Terrestrial laser scanning used to determine the
geometry of a granite boulder for stability analysis purposes. Geomorphology, 106(3-4), pp. 271-277.
- Asensio, E., Vilaplana, J.M. y Abellán, A., 2009. Anàlisi detallada de la susceptibilitat als despreniments a la
cinglera de Castellfollit de la Roca. En: Libro de actas de las Jornades de Recerca i Gestio del Patrimoni
Natural: La vulcanología al Parc Natural de la Zona Volcánica de la Garrotxa, Olot.
- ASPRS LiDAR Committee, 2004. ASPRS Guidelines Vertical Accuracy Reporting for LiDAR Data. Available
at http://www.asprs.org/society/committees/lidar/Downloads/Vertical_Accuracy_Reporting_ for_Lidar
_Data.pdf.
- Ayala- Carcedo, F.J.; Cubillo, S.; Alvarez, A.; Domínguez, M. J.; Laín, L.; Laín, R. y Ortiz, G., 2003. Large
scale rockfall susceptibility maps in La Cabrera Sierra (Madrid) performed with GIS and dynamic
analysis at 1:5,000. Natural Hazards, 30 (3). pp. 325-340
- Baillifard, F., Jaboyedoff, M. & Sartori, M., 2003. Rockfall hazard mapping along a mountainous road in
Switzerland using an empirical approach. Natural Hazards and Earth System Sciences. Vol. 3, 435-442.
- Barton, N., Bandis, S., 1990. Review of predictive capabilities of JRC-JCS model in engineering practise.
Proc.Int.Conference Rock Joints, Balkema publ., Rotterdam, 603-610.
- Bates, R.L., Jackson, J.A., 1987. Glossary of Geology (third edition). American Geological Institute 788p.
- Bauer, A, Paar, G., Kaltenböck, A., 2005. Mass movement monitoring using terrestrial laser scanner for rock
fall management, Proceedings of the First International Symposium on Geo-Information for Disaster
Management, pp. 393–406 Delft, The Netherlands.
- Beraldin, J.A., Blais, F., Boulanger, P., Cournoyer, L., Domey, J., El-Hakim, S.F., Godin, G., Rioux, M.,
Taylor, J., 2000. Real World modelling through high resolution digital 3D imaging of objects and
structures. ISPRS Journal of Photogrametry & Remote Sensing 55, 230-250.
- Biasion, A., Bornaz, L., Rinaudo, F., 2005. Laser Scanning Applications on Disaster Management, in: Geo-
information for Disaster Management, edited by: van Oosterom, P., Zlatanova, S., and Fendel, E. M.,
Springer, Berlin, 19–33.
- Bieniawski, Z.T., 1973. Engineering classification of jointed rock masses. Trans. S. African Inst. Civil
Engineers 15, 335–342.
- Bieniawski, Z.T., 1993. Classification of rock masses for engineering: the RMR system and future trends. In:
Comprehensive Rock Engineering, Hudson J.A. (ed.), Pergamon Press, Vol. 3, 553-573.
- Bitelli, G. Dubbini, M., Zanutta, A., 2004. “Terrestrial Laser Scanning And Digital Photogrammetry
Techniques To Monitor Landslide Bodies”. In: proceedings of the XXth ISPRS congress, Istanbul, Turkey.
Commission V, WG V/2.
- Blanchard, J., Calvet, J., Abellán, A., García, D., Khazaradze, G., Vilaplana, J.M., 2008. Estudio del escarpe
del deslizamiento de Puigcercós mediante láser escáner terrestre. Conca de Tremp, Catalunya. Geotemas
10, 1389-1392.
- Brenner, C., 2006. Aerial laser scanning. International summer school on “Digital Recording & 3D
Modeling”, Aghios Nikulaos, Crete, Greece, 2006, April 24–29.
- Bronshtein, I.N., Semendyayev, Musiol, G., Muehlig, H., 2003. Handbook of Mathematics, 4
th
ed. Springer,
Berlin, 1153 p.
- Budetta, P., 2004. Assessment of rockfall risk along roads. Natural Hazards and Earth System Sciences 4,
71–81.
- Carrara, A., 1983. Multivariate models for landslide hazard evaluation. Mathematical Geology, Vol. XXV:
403-426.
- CETE (Centre d’Etudes Techniques de l’Equipment, Centre de Nice), 1990. Glissement de La Clapiere,
Hydrologie et Cinematique, Donnes Sommaires, Unpublished report.
- Chen, Y., Medioni, G., 1992. Object Modelling by Registration of Multiple Range Images. Image and Vision
Computing 10, 145–155.



Overall References


Antonio Abellán Fernández Page 141
- Copons, R., Altimir, J. Amigó, J., Diaz, A., Vilaplana, J.M., 2000. Eurobloc: un modelo de simulación de
caída de bloques y su máxima adaptación a la realidad. Geotemas, Vol. 1, p.219-222.
- Copons, R., Vilaplana, J.M., 2008. Rockfall susceptibility zoning at a large scale: From geomorphological
inventory to preliminary land use planning. Engineering Geology Vol 102, Issues 3-4, 42-51.
- Copons, R., Vilaplana, J.M., Corominas, J., Altimir, J., Amigo, J., 2005. Rockfall risk management in high
density urban areas. The andorran experience. In Glade, T., et al. Landslide Hazard and Risk. John Wiley
& Sons, pp. 675-698.
- Corominas, J., 1993. Spatial prediction of landslides In: J. Nemec, J.M. Nigg and F. Siccardi (editors).
Prediction and Perception of Natural Hazards, Kluwer Academic Publishers, Dordrecht. pp. 125–132
- Corominas, J., Alonso, E., 1984. Inestabilidad de laderas en el Pirineo catalán. Tipologia y causas.
Inestabilidad de laderas en el Pirineo. Ponencias y Comunicaciones. ETSICCP-UPC, C.1-C.53.
- Corominas, J., Copons, R., Moya, J., Vilaplana, J.M., Altimir, J. and Amigó, J., 2005 Quantitative
assessment of the residual risk in a rock fall protected area, Landslides 2, pp. 343–357.
- Corominas J., Copons R., Vilaplana J.M., Altimir J., Amigó J., 2003. Integrated Landslide Susceptibility
Analysis and Hazard Assessment in the Principality of Andorra. Natural Hazards, November 2003, vol. 30,
iss. 3, pp. 421-435(15).
- Corsini, A., Farina, P., Antonello, G., Barbieri, M., Casagli, N., Coren, F., Guerri, L. Ronchetti, F. Sterzai,
P. and Tarchi, D., 2006. Space-borne and ground-based SAR interferometry as tools for landslide hazard
management in civil protection, International Journal of Remote Sensing 27 (12), 2351–2369.
- Crosta G. B. , F. Agliardi, 2004. Parametric evaluation of 3D dispersion of rockfall trajectories. Natural
Hazards and Earth System Sciences, Vol. 4 (4) 583 – 598
- Crosta, G.B. and Agliardi, F., 2003. Failure forecast for large rock slides by surface displacement
measurements. Canadian Geotechnical Journal, 40(1), 176-191.
- Culebras, J., 2002. Castellfollit de la Roca. Quaderns de la Revista de Girona. Diputació de Girona. 96pp.
- Davis, P. J., 1975. Interpolation and Approximation, Dover publications, New York.
- Demir N., Bayram B., Alkis Z., Helvaci C., Çetin I., Vögtle T., Ringle K., Steinle E., 2004. Laser Scanning for
Terrestrial Photogrammetry, Alternativesystem Or Combined With Traditional System?. In: Proc of the
XXth ISPRS Congress, Istanbul, Turkey, commission V, WG V/2. p. 193-198.
- Derron, M.-H., Blikra, L. H., Jaboyedoff, M., 2005. High resolution digital elevation model analysis for
landslide hazard assessment (Akerneset, Norway), in: Landslides and Avalanches: ICFL 2005 Norway. ed:
Senneset, K., Flaate, K., and Larsen, J. O., Taylor & Francis, London, UK, 101–106.
- Dewez, T., Gebrayel, D., Lhomme, D., Robin, Y., 2009. Quantifying morphological changes of sandy coasts
by photogrammetry and cliff coasts by lasergrammetry. La Houille Blanche (1) 32-37 DOI: 10.1051/
lhb:2009002
- Donville, B., 1973 Geologie néogéne et âges des eruptions volcaniques de la Catalogne Orientale. Thèse
Univ. Paul Sabatier, Toulousse 356pp.
- Dorren, L. 2003. A review of rockfall mechanics and modelling approaches. Progress in Physical Geography
27,1 pp. 69–87
- Dorren, L.K.A., Seijmonsbergen, A.C., 2004. Comparison of three GIS-based models for predicting rockfall
runout zones at a regional scale. Geomorphology, Vol 56, Issues 1-2, pp.49-64
- Douglas, 1980. Magnitude frequency study of rockfall in Co. antrim, Northern-Ireland. Earth Surface
Processes and Landforms, 5(2), pp. 123.
- Dunning, S.A., Massey, C.R. Rosser, N.J., 2009. Structural and geomorphological controls on landslides in
the Bhutan Himalaya using Terrestrial Laser Scanning. Geomorphology, Vol. 103, Issue 1, 17-29.
- Fell, R., Corominas, J., Bonnard, C., Cascini, L., Leroi, E., Savage, W.Z., 2008. Guidelines for landslide
susceptibility, hazard and risk zoning for land use planning. Engineering Geology 102, 85-89.



APPENDIX 1


Page 142 PhD thesis
- Feng, Q., Sjögren, P., Stephansson, O., Jing, L., 2001. Measuring fracture orientation at exposed rock faces
by using a non-reflector total station. Engineering Geology 59, 133-146.
- Fernández, M., Vilaplana, J. M., 2004. Factores meteorológicos que desencadenan desprendimientos de
rocas en Vall de Núria. In: G. Benito and A. Díez Herrero, Editors, Riesgos Naturales y Antrópicos en
Geomorfología vol. II, SEG - CSIC Centro de Ciencias Medioambientales, Madrid, pp. 389–401
- Fernández, O., 2004. Reconstruction of geological structures in 3D: An example from the Southern Pyrenees.
PhD Thesis, Universitat de Barcelona, Spain.
- Fernández, O., 2005. Obtaining a best fitting plane through 3D georeferenced data. Journal of Structural
Geology, 27, 855-858
- Fiani, M. and Siani, N., 2005. Comparison of terrestrial laser scanners in production of DEMs for Cetara
tower. Proceedings CIPA XX International Symposium, Torino.
- Frattini, P., Crosta, G., Carrara, A., Agliardi, F., 2008. Assessment of rockfall susceptibility by integrating
statistical and physically-based approaches. Geomorphology 94, 419-437.
- Fukuzono T., 1990. Recent studies on time prediction of slope failure. Landslide News 4:9–12.
- Fukuzono T., 1985. A new method for predicting the failure time of a slope. In: Proceedings of the fourth
international conference and field workshop on landslides. Tokyo: Japan Landslide Society; p. 145–50.
- Giani, G.P. Giacomini, A. Migliazza, M. and Segalini, A., 2004. Experimental and Theoretical Studies to
Improve Rock Fall Analysis and Protection Work Design, Rock Mech. Rock Eng. 37 (5), 369–389
- Glawe, U., Lotter, M., 1996. Time prediction of rock slope failures based on monitoring results. Senneset, K.
(ed): Landslides, Proceedings of 7th International Symposium on Landslides, 1551-1555, A. A. Balkema,
Rotterdam.
- Glawe, U., Zika, P., Zvelebil, J., Moser, M., Rybti, J., 1991. Time prediction of a rock fall in the Carinthian
Alps. Q. J. Engng. Geology, 26,18S-192.
- Goodman, R.E., Bray, J.W., 1976. Toppling of rock slopes. In: American Society Civil Engineers. Rock
Engineering for Foundations and Slopes; Proceedings of a Specialty Conference, Vol. 2, New York. 201-
233.
- Guinau, M., García-Sellés, D., Santana, D. Mavrouli, O. Corominas, J., Suriñach, E., accepted.
Caracterización automática de discontinuidades en laderas rocosas mediante la aplicación del láser
escáner terrestre (TLS). Ed: Alonso, E., Gili, J.M., Corominas, J., Hürlimann, M., Ledesma, A. Libro de
actas del VII Simposio Nacional sobre Taludes y Laderas inestables, Barcelona, 27-30 Octubre 2009.
- Günter, A., 2003. Slopemap: programs for automated mapping of geometrical and kinematical properties of
hard rock hill slopes, Computer & Geosciences 29/7, 865-875.
- Günther, A., Carstensen, A., Pohl, W., 2004. Automated sliding susceptibility mapping of rock slopes. -
Natural Hazards and Earth System Sciences, Vol. 4, 95-102.
- Guzzetti F., Reichenbach P. & Ghigi S., 2004. Rockfall hazard and risk assessment in the Nera River Valley,
Umbria Region, central Italy. Environmental Management, Vol. 34:2, 191-208
- Guzzetti, F., 2000. Landslide fatalities and the evaluation of landslide risk in Italy. Engineering Geology 58,
89–107.
- Hancock, J., Langer, D., Hebert, M., Sullivan, R., Ingimarson, D., Hoffmann, E., Mettenleiter, M., Fröehlich,
C., 1998. Active laser radar for high-performance measurements. Proceedings of the 1998 IEEE
International Conference on Robotics and Automation, Vol.2, 1465-1470.
- Hartlen, J., Viberg, L., 1988. Evaluation of landslide hazard. In: Bonnard, C. (editor), 5th International
Symposium on Landslides, Lausanne 2, pp. 1037–1058.
- Herda, H.H.W., 1999. Strike standard deviation for shallow-dipping rock fracture sets. Rock Mechanics and
Rock Engineering, 32 (4), 241-255.
- Hoek, E., 2000. Analysis of rockfall hazards. Rock Engineering: course notes, p. 115-136. Rocscience.
Available at <www.rocscience.com>.



Overall References


Antonio Abellán Fernández Page 143
- Hungr, O, Evans S.G., Hazzard J., 1999. Magnitude and frequency of rock falls and rock slides along the
main transportation corridors of southwestern British Columbia. Canadian Geotechnical Journal 36:
224–238.
- Hunter, G., Pinkerton, H., Airey, R., Calvari, S., 2003. The application of a long-range laser scanner for
monitoring volcanic activity on Mount Etna. Journal of Volcanology and Geothermal Research 123, 203-
210.
- Ibsen, M.L. and Brunsden, D., 1996. The nature, use and problems of historical archives for the temporal
occurrence of landslides, with specific reference to the south coast of Britain, Ventnor, Isle of Wight.
Geomorphology Vol. 15, Issues 3-4: Landslides in the European Union, 241-258.
- Ingensand, H., 2006. Metrological aspects in terrestrial laser-scanning technology. In Proceedings of the 3rd
IAG/12th FIG Symposium, Baden, Austria.
- Institut Cartogràfic de Catalunya (ICC), Institut Geològic de Catalunya (IGC) and Parc Natural de la Zona
Volcànica de la Garrotxa (PNZVG), 2007. Carta vulcanològica de la zona volcànica de la Garrotxa
[cartographic material]. Scale: 1:25.000. 93 x 96 cm. Edited by ICC.
- Jaboyedoff, M., Metzger, R., Oppikofer T., Couture, R., Derron, M.-H., Locat, J., Turmel, D., 2007. New
insight techniques to analyze rock-slope relief using DEM and 3D-imaging cloud points: COLTOP-3D
software. In: Eberhardt, E., Stead, D., Morrison, T. (Eds.) Rock Mechanics: Meeting Society's Challenges
and Demands (Vol. 2). Taylor & Francis, pp. 61-68.
- Jaszczak, P., 2006. Terrestrial Laser Scanning for landslides deformation monitoring. MsC thesis. Institute
of Geomatics. 39pp.
- Krautblatter M, Dikau R. 2007. Towards a uniform concept for the comparison and extrapolation of
rockwall retreat and rockfall supply. Geografiska Annaler 89A: 21–40.
- Lan, H., Martin, C. D., Lim, C.H., 2007. RockFall analyst: a GIS extension for three-dimensional and
spatially distributed rockfall hazard modeling. Comput Geosci 33(2):262–279.
- Lato, M., Diederichs, M., Hutchinson, D.J., Harrap., R., 2009. Optimization of LiDAR scanning and
processing for automated structural evaluation of discontinuities in rockmasses. International Journal of
Rock Mechanics & Mining Sciences 46, 194–199.
- Latter, M., Moser, M., Glawe, U., Zvelebil, J., 1998. Parameters and kinematic processes of spreading of
mountain ridges. Proceedings of 8th International IAEG Congress. Vol.II, 1251-1257, A.A. Balkema,
Rotterdam.
- Lee, S., Ryu, J.H., Lee, M.J., won, J.S., 2006. The application of artificial neural networks to landslide
susceptibility mapping at Janghung, Korea. Mathematical Geology 38, 199–220.
- Lichti D.D., Jamtsho M., 2006. Angular resolution of terrestrial laser scanners. Photogrammetric Record
21(114), 141–160.
- Lichti, D.D., 2007. Error modelling, calibration and analysis of an AM-CW terrestrial laser scanner system,
ISPRS Journal of Photogrammetry and Remote Sensing 61 (5), 307–324.
- Lim, M., Petley, D.N., Rosser, N.J., Allison, R.J., Long, A. J., Pybus, D., 2006. Combined digital
photogrammetry and time-of-flight laser scanning for monitoring cliff evolution. Photogrammetric Record,
20(1), 109-129.
- Lim, M., Rosser, N.J., Allison, R.J., Petley, D.N., 2009. Erosional processes in the hard rock coastal cliffs at
Staithes, North Yorkshire, Geomorphology, doi:10.1016/ j.geomorph.2009.02.011.
- Lindenbergh, R. and Pfeifer, N., 2005. A statistical deformation analysis of two epochs of terrestrial laser
data of a lock, Proceedings of the 7th Conference on Optical 3D Measurement Techniques, Vienna,
Austria, 61-70.
- Lotter M., Moser M., Glawe, U., Zvelebil J., 1998. Parameters and kinematic processes of spreading of
mountain ridges. Proc. 8th Int. Congress IAEG, Vancouver, II, 1251-1257; Rotterdam (Balkema).
- Malamud B.D., Turcotte D.L., Guzzetti F., Reichenbach P., 2004. Landslide inventories and their statistical
properties. Earth Surface Processes and Landforms, Vol. 29, n. 6, 687–711, DOI: 10.1002/esp.1064.



APPENDIX 1


Page 144 PhD thesis
- Mallarach, J. M., Mirabell, E, 1976. Informe Geològic sobre les esllavissades de la tardor de 1976 a la
cinglera basàltica de Castellfollit de la Roca, La Garrotxa. (unpublished material).
- Mallarach, J.M., Riera, M., 1981. Els volcans olotins i el seu paisatge. 62-69 i 192- 200pp. Ed. Serpa.
Girona. 250pp.
- Marques, F. M., 2008. Magnitude-frequency of sea cliff instabilities, Nat. Hazards Earth Syst. Sci., 8, 1161-
1171.
- Marquínez J., Menéndez Duarte R., Farias P., Jimenez M., 2003. Predictive GIS-Based Model of Rockfall
Activity in Mountain Cliffs. Natural Hazards, Vol. 30 (3), 341-360.
- Martí, J., Pujades, A., Ferrés, D., Planagumà, L., Mallarach, J. M., 2000. El Vulcanisme. Guia de camp de
la zona volcànica de la Garrotxa. Guies del Parc Natural de la Zona Volcànica de la Garrotxa,
Departament de Medi Ambient de la Generalitat de Catalunya, Barcelona. 106 pp.
- Maruyama, K., Kozima, S., 1994. A proposal of the warning criterion for the landslide crisis depended on the
observation data of sliding distance. Journal of Japan Landslide Society 31(1), 45-51.
- Mascort, D., Martorell, E., Rovira, A., Montserrat, A., Pujol, D., 2004. Estudi geologic de la zona de
Castellfollit de la Roca. Parc Natural de la Zona Volcánica de la Garrotxa, 30p.
- McKean, J., Roering, J., 2004. Objective landslide detection and surface morphology mapping using high-
resolution airborne laser altimetry. Geomorphology, 57, pp.331–351.
- Monserrat, O., Crosetto, M., 2008. Deformation measurement using terrestrial laser scanning data and least
squares 3D surface matching. ISPRS Journal of Photogrammetry and Remote Sensing, 63(1), 142-154.
- Muller, L., 1968. New considerations on the Vajont slide. Felsmechanik und Ingenieurgeologie 6 (1), 1–91.
- Noferini, L., Pieraccini, M., Mecatti, D., Macaluso, G., Atzeni, C., Mantovani, M., et al., 2007. Using GB-
SAR technique to monitor slow moving landslide. Engineering Geology, 95(3-4), 88-98. doi:DOI:
10.1016/j.enggeo.2007.09.002
- Olivera, C., Redondo, E., Lambert, J., Riera Melis, A., Roca, A., 2006. Els Terratrèmols dels segles XIV i XV
a Catalunya. Generalitat de Catalunya, Institut Cartogràfic de Catalunya, 407p.
- Oppikofer, T., Jaboyedoff, M., Blikra, L. H., Derron, M.-H., 2008b. Characterization and monitoring of the
Åknes rockslide using terrestrial laser scanning, in: Proceedings of the 4th Canadian Conference on
Geohazards: From Causes to Management, edited by: Locat, J., Perret, D., Turmel, D., Demers, D., and
Leroueil, S., Presse de l'Université Laval, Québec, Canada, 211-218.
- Oppikofer, T., Jaboyedoff, M., Blikra, L., Derron, M.-H., Metzger, R., 2009. Characterization and monitoring
of the Åknes rockslide using terrestrial laser scanning. Nat. Hazards Earth Syst. Sci., 9, 1003-1019, 2009
- Oppikofer, T., Jaboyedoff, M., Keusen, H.-R., 2008a. Collapse of the eastern Eiger flank in the Swiss Alps.
Nature Geosciences 1(8), 531-535.
- Optech, 2009. ILRIS-3D Intelligent Laser Ranging and Imaging System, from Optech,
http://www.optech.ca/i3dprodline-ilris3d.htm, access 12 June 2009.
- Palau, A., 1976. Dictamen sobre las causas y remedios de las grietas y desprendimientos en el cingle o
acantilado basaltico de Castellfollit de la Roca. Excmo. ayuntamiento de Castellfollit de la Roca. 35pp.
Girona.
- Pallí, L., Trilla, J., 1976. Estudio de los desprendimientos en el acantilado o “cingle”de Castellfollit de la
Roca. Excma. diputación provincial de Girona. 37pp. Girona.
- Palmström, A., 1995. RMi: A rock mass characterization system for rock engineering purposes. PhD thesis,
Oslo University, Norway, 400p. Available at <www.rockmass.net>.
- Pernito, M.A.E., 2008. Rock mass slope stability analysis based on 3D terrestrial laser scanning and ground
penetrating radar. Enschede, ITC, 2008. 112 p. http://www.itc.nl/library/papers_2008/msc/aes/pernito.pdf
- Petrie and Toth, 2008. Introduction to laser ranging, profiling and scanning. In Topographic Laser Ranging
and Scanning: principles and processing. Edited by: Shan, J., Toth, C. K., CRC Press, Taylor & Francis,
590p.



Overall References


Antonio Abellán Fernández Page 145
- Pfeiffer, T.J Higgins, J.D., Andrew, D.R., Schultz, R.J., Beck, R.B., 1995. CRSP: Colorado Rockfall
Simulation Program, Version 3.0a User’s Manual. Colorado School of Mines.
- Piteau and Associated Limited, 1980. Slope stability analisys for rockfall problems: The computer rockfall
model for simulating rockfall distribution. Rock Slope Engineering, p. 62-68. Department of
transportation, Washington DC.
- Rabatel, A., Deline, P., Jaillet, S., Ravanel, L., 2008. Rock falls in high-alpine rock walls quantified by
terrestrial lidar measurements: A case study in the Mont Blanc area, Geophys. Res. Lett., 35, L10502,
doi:10.1029/2008GL033424.
- Rendón, 2004. Evaluación y zonificación de la peligrosidad geológica por desprendimientos en Vall de
Núria, Pirineos orientales España. PhD thesis, Universitat de Barcelona, Spain, 250p.
- Rendón, A., Vilaplana, J.M, 2004. Propuesta para evaluar y cartografiar la susceptibilidad a los
desprendimientos de ladera naturales. Geotemas 6(1), 359-362.
- Rendón, A., Vilaplana, J.M., Copons, R., 2004. El análisis volumétrico de los depósitos de talud: una etapa
en la evaluación de la peligrosidad por desprendimientos. En: G.Benito & A. Díez Herrero (Eds.),
Riesgos Naturales y Antrópicos en Geomorfología, Vol.II. Madrid: SEG – CSIC Centro de Ciencias
Medioambientales, 413 – 421.
- Rodríguez, X., Abellán, A., Calvet, J. accepted. Detección de desprendimientos mediante láser escaner
terrestre en la zona de Puigcercós (Pallars Jussà). Ed: Alonso, E., Gili, J.M., Corominas, J., Hürlimann,
M., Ledesma, A. Libro de actas del VII Simposio Nacional sobre Taludes y Laderas inestables, Barcelona,
27-30 Octubre 2009.
- Romana, M. 1993. A Geomechanical Classification for Slopes: Slope Mass Rating. In comprehensive rock
engineering, Hudson J.A. (ed.), Pergamon Press, Vol. 3.
- Romana, M., 1988. Practice of SMR classification for slope appraisal. In: Bonnard, C. (Ed.), Proc. 5th
International Symposium on Landslides, Lausanne. Balkema, Rotterdam. Vol 2, 1227-1232.
- Ros, X., Palomar, J., Gaete, R., 1996. Estudi geotècnic del cingle de Castellfollit de la Roca. Parc Natural de
la Zona Volcánica de la Garrotxa (PNZVG). 157 pp.
- Rose, N.D., Hungr, O., 2007. Forecasting potential rock slope failure in open pit mines using the inverse-
velocity method. International Journal of Rock Mechanics and Mining Sciences, 44(2), 308-320.
- Rosser, N.J., Lim, N, Petley, D.N., Dunning, S., Allison, R.J., 2007. Patterns of precursory rockfall prior to
slope failure. Journal of Geophysical Research, Vol. 112, No. F4, doi:10.1029/2006JF000642.
- Rosser, N.J., Petley, D.N., Lim, M., Dunning, S. A., Allison, R.J., 2005. Terrestrial laser scanning for
monitoring the process of hard rock coastal cliff erosion. Quarterly Journal of Engineering Geology and
Hydrogeology, 38(4), 363-375.
- Rouiller, J., Jaboyedoff, M., Marro, Ch., 1998. Pentes instables dans le penique valaisan. Ed V/d/f. Zurich.
239p.
- Rowlands, K.A., Jones, L.D. and Whitworth, M. 2003 Landslide laser scanning: A new look at an old
problem. Quarterly Journal of Engineering Geology and Hydrology 36 (2), 155 - 157.
- Ruiz, A. Kornus, W. Talaya, J., Colomer, J.L., 2004. “Terrain Modeling in an Extremely Steep Mountain: a
Combination of Airborne and Terrestrial Lidar”, In: proc. of the XXth ISPRS congress, Istanbul, Turkey.
Vol. XXXV, part B3 Com. III,
- Saito, M., 1969. Forecasting time of slope failure by tertiary creep. Proceedings 7th International
Conference on Soil Mechanics and Found Engineering 2, 677−683.
- Santanach, P., 1974. Estudi tectònic del paleozoic inferior del Pirineu entre la Cerdanya i el riu Ter. Ed.
Fundació Salvador Vives Casajuana, Barcelona, 133p.
- Schweigl, J., Ferreti, C., Nössing, L., 2003. Geotechnical characterization and rockfall simulation of a slope:
a practical case study from South Tyrol (Italy). Engineering Geology, Vol. 67, p.281-296.
- Scioldo, G., 1991. Rotomap: analisi statistica del rotolamento dei massi. Guida informatica ambientale.
Patron, Milano, 81–84.



APPENDIX 1


Page 146 PhD thesis
- Schneuwly, D.M., Stoffel, M., 2008. Spatial analysis of rockfall activity, bounce heights and geomorphic
changes over the last 50 years - A case study using dendrogeomorphology. Geomorphology, 102 (3-4), pp.
522.
- Sequeira, V., Fiocco, M., Bostrom, G., Gonçalves, J.G.M., 2003. 3D Verification of Plant Design. 25th
ESARDA Symposium on Safeguards and Nuclear Material Management, Stockholm, Sweden. Available at:
<http://sir.jrc.it/sir/Publications/040130-DIV-paper%20ESARDA.pdf>
- Settles, E., Göttle, A., Von Poschinger, A., 2008. Slope Monitoring Methods - A State of the Art Report.
ClimChalp, Interreg III B, Alpine Space, Work package 6, Munich, 165pp.
- Singh, S.West, J., 1991. “Cyclone: A Laser Scanner for Mobile Robot Navigation”, Carnegie Mellon
University, Robotics Institute Technical Report, CMU-RI-TR-91-18.
- Slob, S., Hack, H.R.G.K. and Turner, K., 2002. An approach to automate discontinuity measurements of rock
faces using laser scanning techniques. In: Proceedings of ISRM EUROCK 2002. Ed. by C. Dinid da Gama
and L. Riberia e Sousa. Lisboa, Sociedade Portuguesa de Geotecnia, pp. 87-94.
- Slob, S., Hack, R., 2004. 3D Terrestrial Laser Scanning as a new Field Measurement and Monitoring
Technique. In: Hack, R., Azzam, R., Charlier, R. (eds), Engineering Geology for Infrastructure Planning
in Europe: A European perspective. Springer, Berlin. Lecture notes in Earth Sciences, n.104, pp179-189.
- Soeters, R. and Van Westen, C. J.: Slope Instability recognition, analysis and zonation. In Landslides:
Investigation and mitigation. Special report 247, Transportation Research Board. National Research
Council, Washington D.C., pp. 129-177, 1996.
- Soudarissanane, S., Lindenbergh, R. and Gorte, B. Reducing the error in terrestrial laser scanning by
optimizing the measurement set-up. Proc. of the XXI Congress, The International Society for
Photogrammetry and Remote Sensing, ISPRS2008, Vol. XXXVII, Commission V, 3-11 July 2008, Beijing,
China, 615-620, 2008.
- Staiger, R., 2003. Terrestrial laser scanning technology, systems and applications. 2nd FIG regional
Conference, Morocco. Available at <http://www.fig.net/pub/morocco/proceedings/TS12/TS12_3_
staiger.pdf>
- Stoffel, M., Schneuwly, D., Bollschweiler, M., Lièvre, I., Delaloye, R., Myint, M., Monbaron, M., 2005.
Analyzing rockfall activity (1600–2002) in a protection forest – a case study using dendrogeomorphology.
Geomorphology 68, 224–241.
- Sturzenegger, M., Stead, D., 2009. Quantifying discontinuity orientation and persistence on high mountain
rock slopes and large landslides using terrestrial remote sensing techniques. Nat. Hazards Earth Syst.
Sci., 9, 267–287.
- Sturzenegger, M., Stead, D., Froese, C., Moreno, F. Jaboyedoff, M., 2007. Ground-based and airborne
LiDAR for structure mapping of a large landslide: the Frank Slide.In: Eberhardt, E., Stead, D., Morrison,
T. (editors). Proceedings of the first Canadian US rock mechanics symposium, vol.2. London:
Taylor&Francis, p.925–32.
- Tarchi, D. Antonello, G. Casagli, Farina, P., Fortuny-Guasch, J., Guerri, L., Leva, D., 2006. On the use of
ground-based SAR interferometry for slope failure early warning: the Cortenova rock slide (Italy). In
Landslides: Risk Analysis and Sustainable Disaster Management, part IV, 337-342. Ed: Sassa, K.,
Fukuoka, H., Wang, F., Wang, G. 385p Springer, Berlin.
- Teixeira, S.B., 2006. Slope mass movements on rocky sea-cliffs: a power-law distributed natural hazard on
the Barlavento Coast, Algarve, Portugal. Continental Shelf Research 26, 1077–1091.
- Terzaghi, K., 1950. Mechanism of landslides. In: Paige, S. (Ed.) Application of Geology to Engineering
Practice (Berkeley Volume). Geological Society of America, Washington D.C., pp. 83–123.
- Teza, G., Atzeni, C., Balzani, M., Galgaro, A., Galvani, G., Genevois, et al., 2008. Ground-based monitoring
of high-risk landslides through joint use of laser scanner and interferometric radar. International Journal
of Remote Sensing, 29 (16), 4735-4756.
- Teza, G., Galgaro, A., Zaltron, N., Genevois, R., 2007. Terrestrial laser scanner to detect landslide
displacement fields: a new approach. International Journal of Remote Sensing, Volume 28, Issue 16,
pages 3425 – 3446.



Overall References


Antonio Abellán Fernández Page 147
- Van Asch T. W. J., Malet, J-P., Van Beek, L. P. H., Amitrano, D., 2007. Techniques, issues and advances in
numerical modelling of landslide hazard. Bulletin de la Societe Geologique de France, 178, pp. 65-88
- Van Westen, C.J., Van Asch, T.W.J., Soeters, R., 2006. Landslide hazard and risk zonation—why is it still so
difficult? Bull Engineering Geology and Environment 65, 167–184.
- Varnes, D. J., 1978. Slope movement types and processes. In: Schuster, R.L., Krizek, R.J. (Eds.), Special
Report 176: Landslides: Analysis and Control. TRB, National Research Council, Washington, D.C., pp.
11–33.
- Varnes, D.J., 1984, Landslide Hazard Zonation: a Review of Principles and Practice. Commission on
Landslides of IAEG, UNESCO, Paris, Natural Hazards, 3, 61 p.
- Vibert C., Arnould, M., Cojean, R., Le Cleach, M., 1988. Essai de prevision de rupture d’un versant
montaineaux. In: Bonnard C, editor. Proceedings of fifth international symposium on landslides,
Lausanne, vol. 1. Rotterdam: A.A. Balkema, 789–92.
- Vidal, L.M., 1881. Nota acerca de los hundimientos ocurridos en la Cuenca de Tremp (Lérida) en Enero de
1881. Boletin de la Comisión del Mapa Geológico de España, Tomo VIII, 113-129.
- Vilajosana, I., Suriñach, E., Abellán, A., Khazaradze, G., Garcia, D., Llosa, J., 2008. Rockfall induced
seismic signals: case study in Montserrat, Catalonia, Nat. Hazards Earth Syst. Sci., 8, 805-812.
- Voegtle, T., Schwab, I., Landes, T., 2008. Influences of different materials on the measurement of a
Terrestrial Laser Scanner (TLS). Proc. of the XXI Congress, The International Society for
Photogrammetry and Remote Sensing, ISPRS2008, Vol. XXXVII, Commission V, 3-11 July 2008, Beijing,
China, 1061-1066.
- Voight, B., 1989. A relation to describe rate-dependent material failure. Science 243, 200–203.
- Wyllie, D.C., Mah, C.W., Hoek, E., 2004. Rock Slope Engineering: Civil and Mining. Taylor & Francis,
431p.
- Zvelebill, J., Moser, M., 2001. Monitoring Based Time-prediction of rock falls: Three Case-Histories. Phys.
Chem. Earth (B), Vol. 26, No. 2, 159 – 167.



















APPENDIX 1


Page 148 PhD thesis

































Appendix 2
List of tables




APENDIX 2


Page 128 PhD thesis



























LIST OF TABLES


Antonio Abellán Fernández











List of tables


- Table 1.1. Summary of the main characteristics of the different pilot study areas. A detailed analysis
(except for risk and exposure assessments) is discussed in subsequent sections (see Chapters 2, 4 and 6).
- Table 2.1: Terrestrial Laser Scanner Ilris3D technical specifications (www.ilris_3D.com).
- Table 2.2. Data acquisition using TLS-Ilris3D.
- Table 2.3: Characteristics of Digital Elevation Model generated using the point cloud acquired with the
TLS-Ilris3D
- Table 4.1. Main geological and geo-mechanical characteristics of the stratigraphic sequence.
- Table 4.2. Historical record of rockfalls in the 30 years prior to our study.
- Table 4.3. (a) Long-term results based on the historical record and population surveys. Time span: 50
years prior to this study; (b) Short-term results based on a sequential comparison of TLS datasets. Time
span: 22 months.
- Table 4.4: Combination of short and long-term approaches: TLS monitoring campaign, historical record
and population surveys. This table attempts to quantify the recurrence and volume of the failures
rockfalls in the last 50 years. The calculation of a “probability of occurrence” is beyond the scope of this
research.
- Table 5.1. (a) TLS measurements. (b) Calliper measurements.
- Table 6.1 Rainfall vs. rockfall frequency
- Table 7.1 Summary of the results in each Publication (Chapters 2, 4, 5 and 6). These results concern the
different rockfall stages: pre-failure deformation, detachment and evolution of the phenomenon.
- Table 7.2. Areas with precursory deformation detected in 2007-2008.
- Table 8.1. New areas with precursory deformation detected during fieldwork carried out after the
acceptance of Publication D. These new areas confirmed the possibility of early precursory deformation
detection.



APENDIX 2


Page 128 PhD thesis














































Appendix 3
List of figures




APPENDIX 3


Page 154 PhD thesis










List of figures


Antonio Abellán Fernández Page 155










List of figures


Chapter 1
- Fig. 1.1. Number of publications indexed by the SCI Expanded (ISI Web of Knowledge,
www.isiwebofknowledge.com). Analysed period from 2000 to 2008. Keywords (in topic) = (laser scan*
terrest*) or (laser rockfall); Subject Areas = geography or geology; Current year (2009) is not included in
this figure due to an incomplete record.

Chapter 2
- Fig. 2.1. Study area in Vall de Núria, Eastern Pyrenees, Spain. Pilot area is enclosed by dotted line.
- Fig. 2.2. Rack railway and recent rockfalls in the study area. (a) Event B: 2003/04/04. (b) Event C:
2003/06/15. Fig.3. Close up (3a) and front view (3b, from www.ilris-3D.com) of long range Terrestrial
Laser Scanner Ilris-3D
- Fig. 2.4. Methodology used in our study. Step 5 is currently under construction (see section 2).
- Fig. 2.5. Digital Elevation Model (a) and photographs of the study area seen from the stations E2 (b), E1(c)
and E3(d). A total of 8 scans were performed from each of the stations: E1 (E11, E12), E2 (E21, E22, E23,
E24) and E3 (E31, E32). The scanned areas are enclosed by dotted lines.
- Fig. 2.6. (a) Point cloud acquired with TLS Ilris-3D. Selected area is enlarged in Fig. 2.6c. (b) Photography
from station E3. Fig. 6c is enclosed with discontinuous line. (c) and (d) different views of the point cloud.
Selected area is enlarged in Fig. 2.6d.



APPENDIX 3


Page 156 PhD thesis
- Fig. 2.7. Photography (left) and identification of the detachment area in the point cloud acquired with the
TLS-Ilris3D (right).
- Fig. 2.8. (a) 3D point cloud of the detachment area of event C (2003/06/15). Selected area is enlarged in fig.
2.8b. (b). 3D point cloud allows us to model joint geometry: J1, J2 and J3. (c) Photography of the same
area. (d) Stereographic projection of the joints: J1, J2, J3 and direction of movement of the wedge
between J1 and J2.
- Fig. 2.9. Vertexes coordinate and planes orientation of the detached block of event C. DD: dip direction; DA:
dip angle.
- Fig. 2.10. Contour lines (equidistance: 5 m.) plotted using DEM 1 (a) and DEM 2 (b): DEM 2 was obtained
using digital topographic map 1:5.000. DEM 1 was generated using point clouds acquired with TLS ilris-
3D. Despite the fact that both DEM have s similar resolution (6 m), DEM 1 is more accurate than DEM 2.
Coordinates are local.
- Fig. 2.11. Comparison of trajectories observed in the field (solid line) with simulations undertaken using
DEM 1 (dashed line) and DEM 2 (dotted line) for events A and B. The highest accuracy is achieved with
DEM 1.

Chapter 3
- Fig. 3.1 Basic operation of a laser rangefinder (from Petrie and Toth, 2008).
- Fig. 3.2 Shape of the complete waveform of the returned pulse. The first and last pulse can be seen in this
picture (from Petrie and Toth, 2008; original figure by Brenner, 2006).
- Fig. 3.3 Data acquisition using a TLS ILRIS-3D (Optech)

Chapter 4
- Fig. 4.1. (a) Location of the village and the basaltic formation at Castellfollit de la Roca (Garrotxa Volcanic
Field, NE Spain); (b) Perspective view with indication of the scanned rockface; (c) Synthetic stratigraphic
sequence modified from Palli and Trilla (1976). Thickness ~45m. See Table 1 for a description of the
geological levels (from A to I); (d) Panoramic view of the North face of the basaltic formation at
Castellfollit de la Roca. Houses located on the edge of the cliff are visible.
- Fig. 4.2 Type of rockfall according to its volume: (a) detachment of single columns (volume below 1.5 m
3
);
(b) detachment of a group of columns (volume from 1.5 to 150 m
3
); (c) likely slab failure (volume higher
than 150 m
3
).
- Fig. 4.3 Images of the aligned TLS datasets: (a) left view: “-X axis”; (b) zenithal view: “+Z axis”; (c)
frontal view: “–Y axis”.
- Fig. 4.4: Assessment of the quality of the rock face modelling. (a) Variance of the population assessed for
different densities of points through the percentiles of the error of the model comparison; (b) Probability
distribution of the error of the model comparison using a mean point spacing of 4.7cm. The histogram is
characterized by considerable kurtosis and a poor resemblance to a fitted Gaussian distribution (blue
line). See text for a detailed explanation.
- Fig. 4.5. Comparison of the sequential TLS datasets. Each point of the 3D point cloud is colour-coded in
accordance with the changes recorded during the period of comparison (September vs. December 2006).
The failure of a single basaltic column and small changes in vegetation are visible. Pre and post failure
stages are visible on the left.
- Fig. 4.6. (a) TIN surface of a part of the rock face where events i and ii took place; (b) Comparison of the
March and April 2007 datasets; (c) Comparison of the October 2007 and January 2008 datasets. Each
point is colour-coded according to the changes during the period of comparison. Colour scale indicates
positive differences along the Y direction, e.g. rockfalls. Note that event ii is in the same area as event i.




List of figures


Antonio Abellán Fernández Page 157
Chapter 5
- Fig. 5.1 (a) General overview of the experimental setup (range = 50 m) (b) Zoom of the scanned area
showing the 3 moving objects on the fixed part. (c) Perspective views of the TLS point cloud.
- Fig 5.2. (a) Normalised number of observations (Y axis) for each interval (0.5mm classes) of the RAW data
(plane object). Curves show fitted Gaussian distributions (µ = mean value; σ = standard deviation) for 5,
10, 15, 20 and 25 mm (from left to right, respectively). (b) Mean value of the RAW data displacements vs.
real value of the displacement (calliper) for the three objects (plane, hemisphere and a irregular form).
The error bars represent 1σ standard deviation (7.4, 10.1 and 11.2 mm respectively). Standard deviation
for the plane is significantly lower than for the hemisphere and the irregular form.
- Fig. 5.3. RAW data displacement measurements (7.4mm< σ <11.2mm) for different induced displacements:
(a) 5 mm, (b) 10 mm, (c) 15 mm, (d) 20 mm, (e) 25 mm. The colour scale (in millimeters) indicates
displacements up to 30mm.
- Fig. 5.4. Displacement measurements after the 5×5 closest neighbours averaging for different induced
displacements (σplane = 1.3mm): (a) 5 mm, (b) 10 mm, (c) 15 mm, (d) 20 mm, (e) 25 mm. The colour scale
(in millimeters) indicates displacements of up to 30mm.
- Fig 5.5. (a) Study area: Basaltic cliff at Castellfollit de la Roca. The frame corresponds to Fig. 5.5b (b)
Comparison of TLS models (iii) and (iv) (March and April 2007, respectively) showing the 50 m3 rockfall
event in April 2007. Colour scale from 0 (blue) to 1.5 m (red).
- Fig. 5.6. RAW data displacement measurements prior to the 50m3 rockfall; (a) cumulated displacement for 3
months comparison: 2007.03 vs. 2006.12; (b) cumulated displacement for 6 months comparison: 2007.03
vs. 2006.09. σRAW data = 18.0 mm.
- Fig. 5.7. Filtered displacement measurements prior to the 50m3 rockfall using the 24 Nearest Neighbour
averaging technique. (a) cumulated displacement for 3 months comparison: 2007.03 vs. 2006.12; (b)
cumulated displacement for 6 months comparison: 2007.03 vs. 2006.09. A centimetric precursory
displacement is observed in the middle of the figure. σNN Method = 6.4 mm

Chapter 6
- Fig. 6.1. Picture of the study area showing the rock face and the upper part of the scree deposit. Ruins of the
old village of Puigcercós can be seen at the top of the slope.
- Fig. 6.2. Data acquisition performed with the TLS. The final 3D model of the study area was generated by
aligning the datasets. The contact between rock face and the scree deposits is indicated by a black line. (a)
frontal view: “–Y axis”; (b) left view: “-X axis”; (c) zenithal view: “+Z axis”.
- Fig. 6.3. Main events during the period of study (2007.09.29 – 2008.07.23). Events are chronologically
ordered, from A to E. Sign criteria applied to the differences calculation (Di) is showed on the colour
scale (see text for a complete explanation).
- Fig. 6.4. (a) Bar diagram showing the intervals where the different events occurred (from A to F?). The
intervals with the precursory displacements of the predicted events B and F? are also provided. (b) Daily
rainfall during the period of study (Sant Romà d'Abella, from SMC). The relationship between daily events
and mean daily rainfall is depicted by a star symbol (see text for further explanation).
- Fig. 6.5. Comparison of sequential scans showing the evolution of the rock face. (a) period i; (b) period ii;
and (c) period iii. Each of the new rockfalls is indicated by a white circle.
- Fig. 6.6. Centimetric displacement as a precursory indicator to event B. (a) picture of the studied area
showing the area of comparison. (b) scan comparison in period (i) using the 5x5 filtering technique. A
centimetric displacement is observed at the same area where event B took place in period (ii). The
implications of this precursory displacement can be found in the text.
- Fig. 6.7. Progressive pre-failure deformation in area F. (a) study area and analyzed section. (from b to d)
comparison of datasets: period i, ii and iii, respectively. Colour scale from -0.1 up to 0.1m.



APPENDIX 3


Page 158 PhD thesis
- Fig. 6.8. Bar diagram showing an interpretation of the results: measured vs. interpreted pre-failure
deformation (further explanation is provided in the discussion section).



Chapter 7
- Fig. 7.1. Comparison of the sequential TLS datasets. Each point of the 3D point cloud is colour-coded in
accordance with the changes recorded during the period of comparison (a) Pre- and post-failure stages of
the detachment of a single column are visible; (b) event i; (c) event ii.
- Fig. 7.2. Location and delimitation of the main rockfalls (from A to D) that occurred in Puigcercós during
the TLS monitoring campaign. See Chapter 6 for further information.
- Fig.7.3. Comparison of sequential scans showing the evolution of the rock face. Each new rockfall during
period Piii is indicated by a white circle.
- Fig. 7.4. (from a to c) RAW data displacements for different induced displacements (5, 10, 15 mm,
respectively); (from d to f) displacement measurements after the 5×5 Nearest Neighbour averaging for
similar displacements (5, 10 and 15 mm, respectively). The induced displacements are more accurately
computed after NN averaging.
- Fig. 7.5. Filtered cumulated displacement using the NN averaging technique for a 6-month comparison prior
to the 50 m
3
rockfall. A centimetric precursory displacement is observed in the middle of the figure.
- Fig. 7.6. Location of the areas with precursory deformation (Areas F, G, H). The moving areas are delimited
by a sub-vertical fracture (white line). Area F is enlarged and rotated.

Chapter 8
- Fig. 8.1. Location of the areas with precursory deformation (Areas F, G, H, I, J, K). The moving areas are
delimited by a sub-vertical fracture (white line).




























Appendix 4
Specific terminology






APPENDIX 4


Page 160 PhD thesis






























Specific terminology


Antonio Abellán Fernández Page 161













SPECIFIC TERMINOLOGY

- Alignment: the process of rotating and translating a given dataset to a predefined system of reference. Also
referred to as registration.

- Change detection: the study of the changes or differences that occur during a given portion of time. As
regards the sign criterion, positive values equal a lack of material at a given point, i.e. detachment of the
material. Negative values equal an increase in material, i.e. scree deposits, or a displacement towards the
origin of coordinates, i.e. a pre-failure deformation.

- Geomorphological evolution of a rock face: geometrical changes in the rock face due to a set of process.
These processes are usually connected the erosion and/or slope movements.

- Interval: the portion of time that elapses between successive datasets.

- LIDAR: (Light Detection and Ranging) optical remote sensing technology measures the properties of
scattered light to find range and/or other information about a distant target. This can be an aerial or
Ground-Based (GB) instrument.

- Period: the portion of time that elapses between the first and the current dataset. Not to be confused with the
interval.

- Point cloud: a large quantity of 3D points acquired by the TLS through a rangefinder method, also referred
to as scans or TLS datasets. A point cloud can be considered a “data point cloud” or “reference point
cloud”.




APPENDIX 4


Page 162 PhD thesis

- Precursory indicator: some observed and/or measured sign of activity that precedes an event, e.g. a crack
opening, slope creep, acoustic disturbances, micro-seismicity. The detection of a precursory deformation
is the basis for the spatial prediction of rockfalls in our study.

- Precursory (or pre-failure) deformation: surface movements preceding a failure in the slope. In our study,
this deformation ranged from several mm to up to a few decimetres. The detection of this precursory
indicator is the basis for the spatial prediction of rockfalls. Also referred to as slope creep.

- Range: the distance between the TLS and the portion of the terrain for which the coordinates are acquired.

- Rate of retreat: the mean value of detached material (distance) in a given slope for a given period of time (e.g.
mm·year
-1
). The detachment of material could be caused by slope movements, e.g. rockfalls, and/or
erosion.

- Rock face: a steep, almost vertical, rocky slope. When facing the sea, it is referred to as a cliff..

- Rockfall early warning system: procedure designed to warn of a potential rock failure through the detection
of a given precursory indicator, e.g. the detection of a precursory deformation.

- Rockfall frequency: the ratio between the number of new events and a given period of time.

- Rockfall: according to the definition given by Varnes (1978), a rockfall is a fragment of rock (a block) that is
detached by sliding, toppling, or falling, drops down a vertical or sub-vertical cliff, proceeds down a slope
by bouncing and flying along ballistic trajectories or by rolling on talus or debris slopes. A rockfall can be
preceded by pre-failure deformation. Also referred to as an event.

- Scan of reference: the first TLS dataset acquired for a certain area. Subsequent datasets are aligned towards
this dataset.

- Sequential comparison of datasets: the process of alignment plus change detection.

- Slope creep: surface movements preceding a failure in the slope. The detection of this precursory indicator is
the basis for the spatial prediction of rockfalls. This concept was defined by Terzaghi (1950). Also referred
to as pre-failure deformation and/or precursory deformation.

- Spot Dimension: the projection of the laser beam on the ground or target area.

- TLS: Terrestrial Laser Scanner, which is also known as a Ground-Based LIDAR. This instrument was used in
the study to accurately obtain the 3D coordinates of million of points in each dataset.




Appendix 5




Antonio Abellán Fernández Página i



























Appendix 5
Extended summary in
Spanish






Extended summary in Spanish




Página ii Tesis doctoral



Appendix 5




Antonio Abellán Fernández Página iii




Resumen
extendido





i. INTRODUCCION

i.1 Justificación de la investigación

Las caídas o desprendimientos de rocas pequeñas dimensiones (inferiores a 100 m
3
) son el fenómeno
geomorfológico de mayor frecuencia en laderas rocosas (Copons and Vilaplana, 2008), acantilados
marinos (Lim et al., 2009; Dewez et al, 2009) y taludes antrópicos (Budetta, 2004; Guzzetti et al.,
2004). La presente investigación se ha centrado en el estudio de los desprendimientos de rocas por
los siguientes motivos: (a) los desprendimientos son el movimiento de ladera de mayor velocidad
(Varnes, 1978), por lo que su energía de impacto puede alcanzar valores muy elevados, tal y como se
discute en (Pfeiffer et al., 1995; Agliardi and Crosta, 2003; Dorren and Seijmonsbergen, 2004); (b)
la relación magnitud vs. frecuencia de desprendimientos es de tipo exponencialmente inversa
(Douglas, 1980; Teixeira, 2006; Lim et al., 2009). Es decir, los desprendimientos más frecuentes son
los de menores dimensiones; (c) en las últimas décadas se ha producido un notable incremento de la
ocupación humana en zonas de montaña, por lo que su interacción con el fenómeno natural de los
desprendimientos se convierte en un riesgo importante que debe ser estudiado en profundidad,
específicamente en aquellos casos en los que el valor del riesgo excede el valor de un riesgo
aceptable (Fell et al., 2008).

Las distintos enfoques para estudiar el fenómeno de las caídas de rocas son muy variados, entre ellos
la estimación de la frecuencia (p.e. Malamud et al., 2004; Stoffel et al., 2005; Scneuwly and Stoffel,
2008), evaluación de la susceptibilidad (p.e. Ayala-Carcedo et al., 2003; Günther et al., 2004;
Frattini et al., 2008), de la peligrosidad y/o del riesgo (Hungr et al., 1999; Guzzetti, 2000; Copons et



Extended summary in Spanish




Página iv Tesis doctoral
al., 2005; Corominas et al., 2005), etc. La determinación de la posible localización espacial junto
con la estimación de la ocurrencia temporal de futuros desprendimientos son parámetros de vital
importancia para los estudios de peligrosidad por caídas de rocas (Hartlen and Viberg, 1988;
Guzzetti et al., 2004, entre otros). Sin embargo, la obtención de estos parámetros padece un cierto
grado de incertidumbre en la actualidad, tal y como se discute en varios trabajos (ver, por ejemplo
Van Westen et al., 2006). Esta incertidumbre espacio-temporal es tanto mayor cuanto más grande es
el área estudiada y más reducida la información de la que se dispone.

Durante esta última década, se están produciendo una serie de desarrollos innovadores en el campo
de los sensores remotos. Estos instrumentos permiten adquirir información del terreno con una gran
precisión y resolución espacial, empleando sensores con tecnología láser, óptica y/o radar, montados
sobre instrumentales terrestres, aéreos o de tipo satelital. Dichos avances tecnológicos están
actualmente abriendo un gran abanico de posibilidades para visualizar, interpretar y modelizar
diferentes procesos de la superficie de la Tierra. En lo relativo al fenómeno de estudio de esta
investigación (es decir, desprendimientos o caídas de rocas), estas nuevas tecnologías están
revolucionando los enfoques clásicos de caracterización, monitoreo (monitoring) así como
modelización del fenómeno (e.g. Sturzenegger and Stead, 2009; Lim et al., 2009; Oppikofer et al,
2009). De todos estos avances, la posibilidad de conocer de antemano la localización espacial de
futuros desprendimientos así como enlazar la predicción espacial con la predicción temporal de la
fecha de rotura supone un gran reto en la actualidad. Esta investigación se enmarca en el contexto de
los siguientes trabajos relacionados con el empleo del TLS para la detección, seguimiento y/o
predicción de movimientos de ladera: Rowlands et al. (2003); Bitelli et al., (2004); Bauer et al.,
(2005); Rosser et al., (2005); Biasion et al., (2005); Teza et al., (2007); Rosser et al. (2007 );
Oppikofer et al., (2008a); Dewez et al., (2009); Lim et al., (2009).

Este trabajo empieza presentando un amplio abanico de usos y aplicaciones del TLS en
desprendimientos de rocas (capítulo 2). De estas posibles aplicaciones, se ha seguido investigando
acerca de la evolución geomorfológica de una ladera mediante la adquisición periódica de
información 3D. De este modo, se han detectado dos tipologías de cambios en la morfología del
terreno: por un lado, desprendimientos (capítulos 4 y primera parte del capitulo 6) y por otro lado,
indicadores precursores a la ocurrencia de estos desprendimientos (capítulo 5 y segunda parte del
capítulo 6).



i.2 Objetivos

El propósito principal de esta investigación consiste en mejorar nuestro conocimiento actual sobre
el fenómeno de los desprendimientos en cada una de sus fases, desde la deformación precursora
(pre-failure deformation), separación (detachment) de la ladera hasta su posterior transporte y
depósito, empleando para ello el novedoso instrumental Terrestrial Laser Scanner.

Más específicamente, este propósito principal se ha llevado a cabo mediante el planteamiento de los
siguientes objetivos específicos (entre paréntesis, el capítulo en el que se desarrolla el objetivo
específico):



Appendix 5




Antonio Abellán Fernández Página v


(i) Estudio de los posibles usos y aplicaciones del TLS para el estudio del fenómeno de
desprendimientos. Específicamente, en lo relativo a la mejora de la caracterización
geométrica de laderas rocosas: extracción de planos de discontinuidad y generación
de modelos digitales del terreno para la mejora de simulación de caídas de rocas
(capítulo 2).
(ii) Detección de caídas de rocas mediante una adquisición periódica de información
(monitoreo): obtención de volúmenes y frecuencias de desprendimientos ocurridos
durante el periodo de estudio (capítulos 4 y primera parte del capitulo 6).
(iii) Desarrollar una metodología capaz de detectar deformaciones de orden
centimétrico y/o sub-centimétrico (capítulo 5).
(iv) Aplicar dicha metodología para la detección, mediante un análisis retrospectivo, de
la deformación precursora de aquellos desprendimientos ocurridos durante el
periodo de estudio (capítulo 5).
(v) Aplicar dicha metodología para la predicción espacial de futuros desprendimientos
(segunda parte del capítulo 6).

Esta tesis hace un especial hincapié en la detección de cambios empleando el TLS (objetivos ii. iii y
iv) y su aplicación para la predicción espacial de futuros desprendimientos (objetivo v). Este último
objetivo se considera una de las contribuciones más importantes de la presente investigación. Su
potencial para la localización precisa de volúmenes inestables previamente a la ocurrencia de un
desprendimiento tiene importantes implicaciones en lo relativo a la gestión del riesgo,
específicamente en lo relativo a la posible implementación de esta tecnología en un sistema de alerta
temprana (early warning) sobre amplias extensiones del terreno.



i.3 Áreas de estudio
Las áreas de estudio piloto seleccionadas para esta investigación son Vall de Núria, Castellfollit de la
Roca y Puigcercós (capítulos 2, 4 y 6, respectivamente). Estas zonas de estudio cubren un amplio
espectro de litologías, escenarios geomorfológicos, comportamiento geomecánico, niveles de
peligrosidad, exposición y riesgo, tal y como se muestra en la tabla i. En este punto es importante
comentar que una parte de esta investigación (capítulo 5) se ha realizado en un escenario artificial
que permitió un mejor control de los siguientes parámetros: distancia al objeto, cantidad de
deformación inducida, complejidad morfológica del material, etc. Este ensayo se realizó con el
objetivo de desarrollar una metodología capaz de detectar deformaciones de orden centimétrico y/o
sub-centimétrico (objetivo iii).








Extended summary in Spanish




Página vi Tesis doctoral




VALL DE NÚRIA
(capítulo 2)
CASTELLFOLLIT DE LA ROCA
(Capítulo 4)
PUIGCERCÓS
(Capítulo 6)
Litologías
Gneises y esquistos con
múltiples familias de foliación y
diaclasas.
Coladas de lava basáltica (con
edades comprendidas entre 192.000
y 217.000 años) separadas por un
nivel de paleosuelo y piroclastos.
Depósitos fluviales en la base de los
materiales volcánicos.
Margas, limos, arcillas y
calcarenitas. Estratificación
subhorizontal.
Escenario
geomorfológico
Laderas rocosas muy
escarpadas. Valle de alta
montaña generado por
modelado glacial y fluvial.
Acantilado debido a la incisión del río
Fluviá sobre las coladas basálticas.
El acantilado se encuentra en la
Zona Volcánica de la Garrotxa.
Dimensiones del escarpe: 8000 m
2

(altura máxima 40 metros)
Escarpe de coronación del antiguo
deslizamiento de Puigcercós (1881).
Dimensiones del escarpe: 3000 m
2

(altura máxima 25 metros)
Comportamiento
geo-mecánico y
mecanismo de
rotura
Comportamiento frágil.
Estabilidad controlada por
familias de diaclasas y fallas.
Los planos de discontinuidad
generan cuñas de roca.
Mecanismo de rotura: slide.
Dos coladas de lava separadas por
un nivel de debilidad. Las
discontinuidades controlan la
estabilidad local, especialmente en
los niveles columnares. Mecanismo
de rotura: caída libre (fall) y vuelco
(toppling). Se ha detectado
deformación de orden centimétrico
previa a la ocurrencia de un
desprendimiento de 50m
3
.
Alternancia de materiales con
comportamiento plástico (margas) y
frágil (calcarenitas). La
descompresión debido al
deslizamiento de 1881, generó
grietas verticales subparalelas al
talud. Estas discontinuidades de tipo
extensional controlan el mecanismo
de rotura en la actualidad (toppling).
Se ha detectado deformación de
orden centimétrico previa a la
ocurrencia de varios
desprendimientos.
Peligrosidad
(Volúmenes,
frecuencias y
velocidad)
Volúmenes más frecuentes del
orden de 1 a 10 m
3
. Frecuencia
anual de desprendimientos del
orden de 5-6 eventos
1
. Elevada
peligrosidad debido una
considerable pendiente y altura
de caída. Elevadas energías y
altura de saltos (bouncing).
Longitud considerable de las
trayectorias.
Volúmenes más frecuentes por entre
0.5 y 1.5 m
3
, con una frecuencia
anual de desprendimientos (del
orden de 4 desprendimientos/año)
considerablemente inferior al área de
estudio de Puigcercós. Altura de
caída máx.: 40 metros. Trayectorias
de recorrido reducido. Se han
registrado dos episodios del orden
de 1000 m
3
en los últimos 30 años.
Los volúmenes más frecuentes se
encuentran por debajo de 0.5 m
3
.
Elevadísima frecuencia anual de
desprendimientos (más de 50
desprendimientos durante 2008).
Alturas de caída de entre 10 y 25
metros. Trayectorias de recorrido
reducido.
Exposición
(existencia de
elementos
vulnerable)

Existencia de dos escenarios:
el tren cremallera y los
excursionistas del camino
Queralbs-Núria.


Existencia de dos escenarios: en la
parte superior del acantilado se
encuentra la Población de
Castellfollit de la Roca; en la base
del acantilado no existen elementos
vulnerables de forma permanente.
Zona no poblada en la actualidad: la
antigua población de Puigcercós
cambió de emplazamiento
previamente a la ocurrencia del
deslizamiento de 1881.
Riesgo

Tras varios estudios de
peligrosidad geológica se
construyó un túnel que evita la
exposición al peligro por
desprendimientos. Sin
embargo, sigue existiendo un
cierto riesgo (no cuantificado)
para los excursionistas.

El riesgo es reducido en la parte
inferior. La tasa de retroceso debido
a desprendimientos de rocas es
considerable (ver sección 4.4.3), por
lo que el riesgo se centra en la parte
superior del acantilado (población).


Riesgo muy reducido tanto en la
parte superior como inferior del
escarpe debido a que no existe
apenas exposición al fenómeno.
(1) Promedio de los desprendimientos registrados en todo el valle. Extraído de Fernández y Vilaplana (2004)

Tabla i. Resumen general de las principales características de las áreas de estudio donde se ha llevado a cabo la presente
investigación. Un análisis más detallado de estas características (exceptuando la valoración sobre la exposición y riesgo) se
encuentra en los capítulos 2, 4 y 6 (publicaciones A, B y D, respectivamente).




Appendix 5




Antonio Abellán Fernández Página vii





ii. MATERIAL Y MÉTODO

ii.1 Material

ii.1.1 TLS ILRIS 3D

El instrumental empleado en esta investigación está basado en la tecnología LIDAR (Light Detection
and Ranging). Más específicamente, la adquisición de datos se ha realizado mediante un instrumento
de tipo terrestre denominado Láser Scanner Terrestre (TLS, de sus siglas en inglés Terrestrial
Laser Scanner). El modelo utilizado ha sido un ILRIS3D (acrónimo de Intelligent Laser Ranging
and Imaging System) de la compañía Optech. Dicho instrumento consta de un emisor/receptor de
pulsos láser de longitud de onda infrarroja (1535nm) y un dispositivo de scanner consistente en un
sistema de espejos rotatorios. Puesto que el haz láser se refleja directamente sobre la superficie del
terreno, no es necesaria la existencia de prismas reflectores. El TLS permite la adquisición de las
coordenadas del terreno en un sistema polar (distancia, ángulo horizontal y ángulo vertical), tal y
como se muestra a continuación:

(i) El origen de coordenadas (P
0
=0,0,0) se sitúa en el interior del TLS, es decir, en el punto
de origen de los pulsos láser.

(ii) La distancia (range, ρ) se obtiene midiendo el “tiempo de vuelo” (Time of flight, TOF),
es decir, el tiempo que tarda un pulso láser en recorrer 2 veces (ida y vuelta) la distancia
hasta el punto medido, de acuerdo con la ecuación [1]:
ρ=cTOF/2, donde c=velocidad de la luz
[ecuación 1, Petrie and Toth, 2008]

(iii) Los ángulos horizontal y vertical (θ y ϕ, respectivamente) se obtienen en función de la
resolución angular definida por el usuario, de acuerdo con las ecuaciones [2] y [3]:
θ
i

i-1
+ ∆θρ
i
[ecuación 2, Petrie and Toth, 2008]
ϕ
i

i-1
+ ∆ϕρ
i
[ecuación 3, Petrie and Toth, 2008]




Extended summary in Spanish




Página viii Tesis doctoral
La gran velocidad en la adquisición de datos (2.500 puntos/segundo) permite adquirir una densa
información geométrica del terreno, del orden de varios millones de puntos por cada adquisición,
empleando para ello un tiempo promedio de entre 10 y 25 minutos. Un conjunto de puntos 3D del
terreno se denomina de forma genérica “nube de puntos” (point cloud). La resolución angular es
una medida de la densidad de información, pudiendo obtenerse un espaciado entre puntos de orden
centimetrico e incluso milimétrico. Dicha resolución es función del tamaño de la “huella” del láser
sobre el terreno (S
D
, spot size). A pesar de que el haz láser es una fuente de energía altamente
colimada, la variación del tamaño del spot depende de la divergencia del haz láser y de la distancia
al objeto (Lichti and Jamtsho, 2006), de acuerdo con la siguiente ecuación:

S
D
= ρ·tanα + a [ecuación 4, Petrie and Toth, 2008]

Donde S
D
= tamaño de spot; α = ángulo de divergencia (0.00974º); a= tamaño de spot inicial (12
mm); Nota: Ecuación válida en cualquier sistema de unidades consistente de S
D
y ρ.

La precisión en la medida es función, principalmente, de la distancia de medida, reflectividad del
material y ángulo de incidencia (Lichti, 2007; Petrie and Toth, 2008; Optech, 2009). A pesar de que
la precisión estipulada por el fabricante bajo unas ciertas condiciones ideales es bastante elevada (0.7
cm. a una distancia de 100m, Optech 2009). Sin embargo, el error instrumental es significativamente
superior en laderas naturales (de entre 2 a 4 cm., dependiendo principalmente de la distancia). El
TLS permite adquirir una cuarta variable correspondiente a la cantidad de energía con la que el pulso
láser es recibido. Este parámetro se conoce con el nombre de Intensidad, y es función de los
siguientes parámetros: distancia, ángulo de incidencia, humedad, color y rugosidad del material. Su
valor suele normalizarse posteriormente en una escala 0-255. La medida de la distancia puede
realizarse empleando la primera o última respuesta del pulso enviado (first or last pulse data
acquisition, ver Capítulo 3, Fig.3.2). El alcance máximo del TLS en laderas naturales suele ser del
orden de 600 a 700 metros. Por último, hay que resaltar que el TLS es una tecnología bastante
novedosa y que debido a las características mencionadas anteriormente (datos 3D adquiridos con una
precisión centimétrica y con una gran resolución espacial) está siendo cada vez más empleado en la
caracterización y vigilancia y seguimiento (monitoreo) de taludes y laderas rocosas (a modo de
ejemplo, Bitelli et al., 2004; Biasion et al., 2005; Lim et al., 2006; Rosser et al., 2007; Jaboyedoff et
al., 2007; Sturzenegger et al., 2009; Dunning et al., 2009; Oppikofer et al., 2009; Abellán et al.,
accepted.).


ii.1.2 Adquisición de datos

La adquisición de datos con el TLS en las distintas zonas de estudio comentadas anteriormente
(sección i.3) se ha realizado teniendo en cuenta un amplio espectro de condiciones (laderas naturales
o ensayos experimentales), con distintos valores de densidad de información (espaciado entre puntos
de 0.5cm a 20 cm.), con distinto número de estaciones (de 1 a 5) y distancias a la ladera (de 50 a 870



Appendix 5




Antonio Abellán Fernández Página ix
m) en función de las características específicas de cada una de ellas (accesibilidad, puntos vista
disponibles, dimensiones del área de estudio, etc.). Una descripción más detallada se realiza en cada
una de las publicaciones presentadas en esta tesis (capítulos 2, 4, 5 y 6). A modo de resumen de la
adquisición de datos, cabe resaltar que la geometría de la ladera es tanto más precisa cuanto mayor
es la densidad de información (sección 4.2.4) y menor la distancia desde la que se adquieren los
datos (Petrie and Toth, 2008). Otros parámetros que entran en juego son la reflectividad del material
(Voegtle et al., 2008), morfología de la superficie estudiada (tabla 4.1, capítulo 4) y ángulo de
incidencia (Lichti, 2007). Del mismo modo, la presencia de zonas de sombra (occluded areas) se
puede minimizar adquiriendo los datos desde varias estaciones con distintos puntos de vista. En el
capítulo 2 se ha realizado una única campaña de adquisición de datos con objeto de estudiar las
principales aplicaciones del TLS en laderas rocosas. En las publicaciones sucesivas (capítulos 4, 5 y
6), se ha repetido la adquisición de datos en varias campañas, por lo que ha sido posible estudiar la
variación temporal de la ladera (ver apartado ii.2). En todas las campañas de campo, el archivo
binario generado por el TLS ha sido transformado a un archivo de coordenadas cartesianas
empleando el software Parser (v.4.3.5.4 y anteriores). Posteriormente se ha empleado el software
Polyworks v9.0 para la visualización y tratamiento de la nube de puntos, tal y como se indica en los
apartados sucesivos.



ii.2 Metodología: detección de cambios

ii.2.1 Comparación secuencial de datos del TLS

El empleo del TLS para la detección de cambios en la morfología de la ladera a lo largo del tiempo
ha sido una de las partes más importantes de esta investigación (capítulos 4, 5 y 6). Este enfoque
(approach) fue utilizado de manera pionera por el grupo de investigación del Geography department,
Durham University (Rosser et al., 2005; Lim et al., 2006), grupo con el que se estableció una
colaboración durante el periodo 2006-2007. A continuación se muestra una lista de los pasos
seguidos para llevar a cabo esta comparación (ver sección 4.2):

(i) Primera adquisición de datos empleando el TLS (ver apartado anterior). Esta primera
adquisición constituye lo que de aquí en adelante se denominará “Nube de puntos de
referencia”.

(ii) Construcción de la “Superficie de Referencia” (S
0
): cada uno de los puntos 3D de la
“Nube de puntos de referencia” ha sido definido como un nodo de una superficie TIN
(Triangle Irregular Network). La conexión entre nodos se realiza buscando los 3 puntos
más cercanos, empleando para ello un plano de referencia perpendicular al vector de la
ladera en la parte central del área adquirida con el TLS.

(iii) Repetición de la adquisición de datos: tras un cierto intervalo de tiempo (variable e
función de la zona estudiada y del propósito del estudio) se realiza una nueva



Extended summary in Spanish




Página x Tesis doctoral
adquisición de datos. Cada uno de estos nuevos datos se denominarán de aquí en
adelante “Nubes de puntos de comparación”, y serán numeradas cronológicamente (D
1
,
D
2
, D
3
……. D
n
, del inglés: Data point cloud). De acuerdo con diversos autores (Rosser
et al., 2005; Lim et al., 2006), la adquisición de datos desde la misma estación permite
lograr una mayor precisión en la comparación (ver epígrafe v. Detección de cambios).

(iv) Alineamiento de las D
i
con respecto a S
o
utilizando un algoritmo de cálculo
denominado ICP (Iterative Closest Point). Este algoritmo está basado en la
minimización de la distancia entre puntos de D
i
con respecto a S
o
mediante una función
de mínimos cuadrados. El radio de búsqueda puede ser definido por el usuario. El
método se describe en profundidad en Chen and Medioni (1992).

(v) Detección de cambios: el cálculo de las variaciones en la morfología del terreno entre
sucesivas adquisiciones de datos, se realizó para cada uno de los puntos (P
i
) de las
Nubes de puntos de comparación (D
i
), con respecto a la Superficie de Referencia (S
o
),
de acuerdo con la siguiente ecuación:

Diff
i
= Distancia [D
i
P
0
] - Distancia [S
0
P
0
] [ecuación 5]
Donde Diff
i
= diferencias en cada punto; P
0
= origen de coordenadas = (0,0,0).

El criterio de signos empleado es el siguiente: las diferencias “positivas” (es decir, cuando la
distancia recorrida por el pulso láser es mayor en la superficie original (S
0
) que en las sucesivas
nubes de puntos) equivalen a “falta” de material (es decir, caídas de rocas); por otro lado, las
diferencias “negativas” se corresponden a un “exceso” de material (es decir, depósito) o a un
desplazamiento hacia el origen de coordenadas (como por ejemplo, una deformación precursora a la
ocurrencia de un desprendimiento, ver capítulos 5 y 6). Con objeto de evitar artefactos debido a una
comparación oblicua entre nubes de puntos, la dirección de comparación fue definida a partir del
vector normal de la ladera (escarpe o ensayo experimental) en su parte central. El software
Polyworks v9.0 (módulos ImAlign e ImInspect) fue el software empleado para el alineamiento y la
detección de cambios.



ii.2.2 Detección de pequeños desplazamientos

Este apartado describe la metodología desarrollada en el capítulo 5 para conocer si el error
instrumental del TLS permite detectar cambios de orden centimétrico y/o subcentimétrico. Para ello,
se realizó un ensayo experimental en colaboración con el Institut de Géomatique et d'Analyse du
Risque (IGAR), Université de Lausanne, Suisse. Este ensayo consistía en la simulación de la
deformación precursora mediante el desplazamiento inducido de tres objetos: un plano, una
semiesfera y un cuerpo irregular. Los desplazamientos inducidos en los tres objetos fueron de 0 a 25
mm. en incrementos de 5 mm. entre cada nueva adquisición de datos. Diversos autores (Lindenbergh



Appendix 5




Antonio Abellán Fernández Página xi
and Pfeifer, 2005; Araiba, 2006) discuten como el error en la medida puede reducirse
considerablemente empleando la información de los puntos próximos. Por este motivo la dispersión
de valores de los datos originales (RAW datasets), se promediaron con los valores de los puntos más
cercanos, mediante la técnica de los vecinos próximos (NN: Nearest Neighbour). La búsqueda de
puntos próximos fue optimizada empleando una malla cuadrada. En la sección de resultados se
muestra una comparativa entre los resultados esperables y los resultados medidos empleando ambos
métodos (es decir, los datos originales y los datos promediados).


ii.2.3 Predicción espacial de caídas de rocas.

En nuestra investigación se ha estudiado la existencia de una deformación precursora de orden
centimétrico y/o subcentimetrico previamente a la ocurrencia de un desprendimiento de pequeñas
dimensiones para poder predecir la localización espacial de futuros eventos. Dicha deformación
precursora ha sido estudiada anteriormente en deslizamientos de grandes dimensiones en materiales
dúctiles (volúmenes por encima de 10
6
m
3
, como por ejemplo, Terzaghi, 1950; Saito, 1969; Voight,
1989) así como en grandes desprendimientos rocosos (volúmenes entre 10
4
y 10
7
m
3
, por ejemplo,
Glawe et al., 1991; Zvelebill and Mosser, 2001; Crosta and Agliardi, 2004; Rose and Hungr, 2007).

La posibilidad de detección de cambios de orden centimétrico y/o subcentimétrico, aplicando el
método NN descrito anteriormente a un conjunto de datos adquiridos con el TLS, permitió detectar la
deformación precursora en varios casos reales de caídas de rocas de pequeñas dimensiones (por
debajo de 100 m
3
). Para ello se emplearon dos enfoques: (a) mediante un análisis retrospectivo (back
análisis): estos desprendimientos se produjeron en dos zonas de estudio claramente diferenciadas:
Castellfollit de la Roca (ver sección 4.4) y Puigcercós (ver sección 7.4); (b) Mediante la detección
temprana de esta deformación. De este modo se ha podido predecir la localización espacial de
futuras caídas de rocas en el escarpe rocoso de Puigcercós, tal y como se explica en las diferentes
secciones de esta tesis (ver secciones 6.3.2, 7.4 y 8.4).

















Extended summary in Spanish




Página xii Tesis doctoral












iii. RESULTADOS

Una tabla resumen de los resultados obtenidos en la presente investigación se muestra a continuación
(tabla ii)


PUBLICACIÓN
(capítulo)
DETECCIÓN DE LA
DEFORMACIÓN PREVIA AL
DESPRENDIMIENTO
CARACTERIZACIÓN DEL
DESPRENDIMIENTO
EVOLUCIÓN DEL
FENÓMENO
Public. A
(capitulo 2)
x
Estudio de los planos de
discontinuidad, volúmenes y zona
de salida de desprendimientos
ocurridos durante el periodo de
estudio (análisis retrospectivo)

Mejora de las
simulaciones de caídas
de rocas en base a un
DEM de alta precisión y
de alta resolución.

Public. B
(capitulo 4)
x
Cálculos de volúmenes y
frecuencias de desprendimientos
ocurridos durante el período de
estudio (monitoreo).
Estimación de la tasa
de retroceso.
Public. C
(capitulo 5)
Desarrollo de metodologías para
la detección de la deformación
previa al desprendimiento.
Estudio de la deformación previa
a un desprendimiento en un caso
real (mediante análisis
retrospectivo).
x x
Public. D
(capitulo 6)
Detección de la deformación
previa a la caída (monitoreo).
Predicción de la localización
espacial de futuros
desprendimientos.
Cálculos de volúmenes y
frecuencias de desprendimientos
ocurridos durante el período de
estudio (monitoreo).

Tabla ii. Tabla resumen mostrando los resultados presentados en cada una de las publicaciones. Los resultados se
muestran para cada una de las fases de un desprendimiento, desde la detección de la deformación previa a su
ocurrencia, la caída en sí misma (detachment), así como la evolución posterior del fenómeno y sus implicaciones en
la tasa de retroceso del escarpe.



Appendix 5




Antonio Abellán Fernández Página xiii
iii.1 Principales usos del TLS en caídas de rocas
En esta sección se ha optado por separar los resultados en base a si se ha realizado una única
adquisición de datos o si se ha realizado un seguimiento de la zona de estudio. Por un lado una
única adquisición de datos (capítulo 2) nos ha permitido: (a) obtener las coordenadas y volúmenes
de la zona de salida (sección 2.4.1.1) en base a un análisis retrospectivo; (b) calcular la orientación
de los planos de discontinuidad (sección 2.4.1.2) basándonos en Fernández et al. (2004), y (c)
generar un DEM de alta precisión para mejorar la simulación de caídas de rocas (sección 2.4.2.1).
Por otro lado, una múltiple adquisición de datos nos ha permitido estudiar los cambios que se
producen en la ladera, como por ejemplo caídas de rocas (secciones 4.3 y 6.3.1) y desplazamientos
de pequeñas dimensiones (secciones 5.3 y 5.4). Estos desplazamientos han sido identificados como
indicadores precursores a la ocurrencia de desprendimientos, por lo que han sido empleados para la
predicción de la ocurrencia de nuevas caídas de rocas (secciones 6.3.2 y 7.4). Puesto que el énfasis
de la tesis consiste en la detección de cambios con el TLS, los resultados derivados de una múltiple
adquisición de datos se desarrollarán con mayor profundidad en las secciones sucesivas del resumen
(secciones iii.2, iii.3 y iii.4).



iii.2 Detección de caídas de rocas.
La técnica de detección de cambios (sección 3.2.1) permite obtener la localización, morfología,
magnitud y frecuencia de los desprendimientos ocurridos durante el periodo de estudio. Dos áreas de
estudio piloto han sido analizadas: Castellfollit de la Roca (capitulo 4) y Puigcercós (capítulo 6):

(a) Por un lado, el acantilado de Castellfollit de la Roca ha sido monitoreado durante un
periodo de 22 meses (Marzo 2006 – Enero 2008), obteniéndose los siguientes resultados
(sección 4.3):
- La caída, de manera aislada, de seis columnas basálticas. La magnitud de cada
una de ellas se encuentra por debajo de 1’5 m
3
(Fig. 4.5, capitulo 4).
- La caída de un grupo de columnas en la parte central del acantilado. La
magnitud de este desprendimiento es de 50m
3
(Fig. 4.6, capitulo 4).

(b) por otro lado, los desprendimientos detectados en el escarpe de Puigcercós (capítulo 6)
se corresponden a un periodo de casi 10 meses (01/10/2007 hasta 23/07/2008), obteniéndose
una frecuencia de desprendimientos considerablemente superior a la detectada en la zona
anterior (42 desprendimientos en 10 meses frente a 7 desprendimientos en 22 meses). A
modo de ejemplo de este tipo de resultados, la Figura 6.3 (capítulo 6) muestra los
principales desprendimientos ocurridos en Puigcercós durante el periodo de estudio. Por
último, cabe destacar que en esta zona de estudio se ha encontrado una relación directa entre
la ocurrencia de pequeños desprendimientos y la lluvia promedio (mm./día). Un análisis
más detallado de los resultados en esta zona se encuentra en la sección 6.3.




Extended summary in Spanish




Página xiv Tesis doctoral
iii.3 Detección de pequeños desplazamientos
A continuación se muestran los resultados del ensayo experimental realizado con objeto de
desarrollar metodologías aplicables a la detección de desplazamientos de orden centimétrico y/o
subcentimétrico. Por un lado, la figura 5.3 (ver capítulo 5) muestra los cambios detectados
empleando los datos originales del TLS (RAW data). Por otro lado, la figura 5.3 (ver capítulo 5)
muestra los mismos desplazamientos tras aplicar una técnica de promediado de los puntos cercanos o
vecinos próximos (NN averaging technique). Esta ultima metodología permitió obtener una mejora
considerable en la precisión (σ
NN
= 1.3cm frente a σ
RAW
= 7.2cm). Por este motivo, la citada técnica
fue aplicada a la comparación de los datos previos a la ocurrencia del evento de 50m
3
que tuvo lugar
en Castellfollit de la Roca (Abril 2007, sección 5.4). Mediante un análisis posterior a la fecha de
ocurrencia de este desprendimiento (análisis retrospectivo), se detectó una zona en donde se había
producido una deformación máxima de 4.5 cm. durante un período de 6 meses (ver Figura 5.7). Esta
misma zona se correspondía con la zona en la que se produjo dicho desprendimiento. Por este
motivo, la detección de esta pequeña deformación permitió, tal y como se discute más adelante, su
identificación como indicador precursor de desprendimientos. Este hecho permite la predicción de la
futura zona de caída.


iii.4 Predicción espacial de caídas de rocas.
La existencia de deformación precursora en desprendimientos de pequeñas dimensiones ha sido
probada tanto en Castellfollit de la Roca como en Puigcercós. De estas deformaciones se ha
detectado: (a) dos zonas con deformaciones centimétricas crecientes en el tiempo seguidas de la
ocurrencia de un desprendimiento en cada una de ellas (Evento i y Evento B); (b) tres zonas en
Puigcercós con deformaciones centimétricas crecientes en el tiempo, en donde la ocurrencia de un
desprendimiento es muy probable (Áreas F, H, I). Una tabla-resumen con las características de cada
una de estas zonas se muestra a continuación (tabla iii)


Deformación precursora Nombre Zona de
estudio
Volumen
(m
3
) Extensión
máx. (cm.)
Periodo
(días)
Ocurrencia
desprendimiento
Predicción /
Análisis
retrospectivo
Evento i
Castellfollit
de la Roca
50 4.5 P
3 a 4
(182) SI
Análisis
retrospectivo
Área B
Puigcercós 2 6 P
i
(57) SI Predicción
Área F
Puigcercós - 6 P
i a v
(546) - Predicción
Área G
Puigcercós . 6 P
i a v
(546) - Predicción
Área H
Puigcercós . 4 P
i a v
(546) - Predicción
Tabla iii. Áreas con deformación precursora creciente detectada en las distintas zonas de estudio piloto.


Las deformaciones precursoras detectadas en Puigcercós coinciden en las siguientes características:
(a) la existencia de una fractura subvertical que delimita el bloque en movimientos del resto del



Appendix 5




Antonio Abellán Fernández Página xv
escarpe; (b) la existencia de un desplazamiento que aumenta su valor con la altura del bloque en
movimiento. De acuerdo con diversos autores (Muller 1968; Goodman and Bray, 1976; Varnes,
1978), este movimiento es característico de un mecanismo de rotura de tipo vuelco o toppling, es
decir, un movimiento rotatorio a lo largo de una base fija; (c) se han detectado pequeños
desprendimientos en la parte superior del bloque en movimiento.
















iv. DISCUSIÓN

iv.1 Análisis crítico de la investigación

iv.1.1 Fortalezas y debilidades del instrumental

- Las fortalezas del instrumental se enumeran a continuación: (a) elevada precisión puntual: 0.72cm.
a una distancia de 50m (capitulo 5); (b) elevada densidad de información, pudiendo lograrse un
espaciado de puntos de orden subcentimétrico; (c) rápida adquisición de datos (2500 puntos/seg.); (d)
posibilidad de cubrir grandes extensiones de terreno, como por ejemplo, escarpes, acantilados y
taludes con extensiones superiores (por ejemplo 8·10
3
m
2
en Castellfollit, ver capitulo 4); (e)
considerable alcance máximo en laderas naturales, en función de la reflectividad del material
(usualmente 600-700m; máximo alcance: 870 m., ver capitulo 2); (f) no es necesaria la existencia de
prismas reflectores intermedios (por este motivo, el TLS permite el estudio de zonas inaccesibles y
peligrosas).

- Las principales debilidades encontradas son: (a) es necesario una visión directa entre el TLS y la
ladera a estudio, es decir, sin la presencia de objetos intermedios (vegetación, cables, etc); (b) la
longitud de onda empleada por el TLS ILRIS3D no permite la adquisición de datos del terreno



Extended summary in Spanish




Página xvi Tesis doctoral
cuando existe nieve o agua en la ladera; (c) nuevos modelos de TLS permiten un mayor alcance
máximo. Por ejemplo, el ILRIS 3D extended range permiten obtener un alcance máximo un 40%
más elevado, (Optech, 2009); (d) la detección de cambios en una dirección determinada de antemano
(perpendicular a la ladera) es una simplificación que se ha considerado asumible en este estudio.
Otras metodologías permiten la obtención del campo de desplazamientos en 3D mediante la
aplicación de la matriz de roto-traslación a zonas concretas de la nube de puntos (ver por ejemplo
Monserrat and Crosetto, 2008; Oppikofer et al., 2009); (e) la detección de desplazamientos
subcentimétricos en laderas naturales queda enmascarado por el error instrumental, incluso tras la
aplicación de la metodología de Nearest Neighbours (ver capitulo 5); (f) la limitación más
importante es desde un punto de vista temporal: la frecuencia de desprendimientos en Castellfollit de
la Roca no es lo suficientemente elevada para poder caracterizar sus relaciones Magnitud-Frecuencia
para un periodo de tiempo reducido (22 meses). Por ello fue necesario completar los datos obtenidos
con el TLS con los obtenidos mediante el análisis del registro histórico (ver capítulo 4); (g)
finalmente, el intervalo de tiempo existente entre adquisiciones de datos sucesivas fue demasiado
elevado: una adquisición de datos más frecuente nos hubiera permitido analizar la variable temporal
con una mayor precisión.


iv.1.2 Comparación con otros trabajos

El número de publicaciones en las que se emplea el TLS en el campo de las ciencias de la Tierra ha
crecido de un modo exponencial en estos últimos años (ver figura 1.1). A nivel nacional, la presente
investigación abre una línea de investigación innovadora. En este sentido, tan solo hemos
encontrados dos publicaciones en revistas de impacto en las que se aplica el TLS para el estudio de
movimientos de ladera (Monserrat and Crosetto, 2008; Armesto et al., 2009). Respecto a las
publicaciones encontradas en bibliografía a nivel internacional en lo relativo a la detección de
desprendimientos empleando el TLS, cabe comentar que el capitulo 4 y la primera parte del capitulo
6, consisten en la aplicación de una metodología desarrollada previamente (ver Roser et al. 2005;
Lim et al. 2006) a dos nuevas zonas de estudio (Castellfollit de la Roca y Puigcercós). La primera de
ellas se caracteriza por una frecuencia de desprendimientos considerablemente inferior a estudios
previos realizados por diversos autores (como por ejemplo Roser et al. 2005; Lim et al. 2009, Dewez
et al. 2009), e incluso a la frecuencia detectada en Puigcercós (ver capitulo 6). Por estos motivos,
nuestros intentos por estimar la tasa de retroceso del acantilado hubieron de completarse con
estudios históricos (ver secciones 4.1.2 y 4.4.3);

La principal contribución de esta tesis es el empleo del TLS para la detección de pequeños
desplazamientos (capitulo 5) y su aplicación para la predicción de caídas de rocas (segunda parte del
capitulo 6). La predicción de la ocurrencia de movimientos de ladera encontrados en bibliografía se
basan, generalmente, en una estimación temporal de la fecha de rotura a partir de puntos concretos
de vigilancia (como por ejemplo, Saito, 1969, Voight, 1989, Glawe et al., 1991; Lotter et al., 1998;
Zvelebil and Moser, 2001; Crosta and Agliardi, 2004; Rose and Hungr, 2007). Esto puede ser
debido a que los sensores de medición empleados en estos trabajos ha sido de tipo puntual (como por



Appendix 5




Antonio Abellán Fernández Página xvii
ejemplo extensómetros, estación total, etc.), por lo que no son capaces de adquirir información de
forma masiva sobre grandes áreas. Dichos motivos pueden haber contribuido a que la predicción
espacial de futuros desprendimientos sea una línea de investigación muy reciente. Excepto por la
investigación realizada por Rosser et al., (2007), la bibliografía al respecto de la predicción de la
localización espacial de futuros desprendimientos es muy escasa. Dicho trabajo se basa en la
detección de la ocurrencia de pequeños desprendimientos, identificando este fenómeno como
indicador precursor de caídas de rocas de mayores dimensiones. Se ha encontrado una gran densidad
de desprendimientos de orden decimétrico en una de las zonas en la que se había detectado
previamente una deformación precursora creciente (Área H, Puigcercós), por lo que ambos
indicadores precursores pueden darse de manera simultánea.

El concepto de predicción espacial empleado en esta investigación es significativamente distinto a la
predicción espacial estudiada por otros autores (Brabb et al., 1972; Carrara et al, 1983; Corominas,
1993; Lee et al., 2006). Si bien ambos enfoques pueden ser complementarios en distintas escalas de
trabajo, el concepto de predicción espacial encontrado en bibliografía suele estar referido a la
susceptibilidad a la caída o “probabilidad de ocurrencia”. Esta suele calcularse en base al
procesamiento de la información espacial contenida en una base de datos (como por ejemplo
orientación de los planos de discontinuidad, orientación de la ladera, litología, etc.) empleando para
ello diversas técnicas (por ejemplo, Bayesian probability, neural networks, likelihood ratios, fuzzy
sets, etc.), con el objetivo de encontrar zonas con las mismas características que aquellas en las que
se ha producido un movimiento de ladera en el pasado. Las principales diferencias con respecto a
nuestra propuesta son: (a) en primer lugar, mediante la detección de la deformación precursora, se
delimita de un modo mucho más preciso la zona en donde se está “gestando” un desprendimiento; (b)
en segundo lugar, no se realiza únicamente una predicción espacial, sino que es posible considerar la
variable temporal en la predicción de ocurrencia del desprendimiento: una vez que la deformación
precursora es detectada y ésta va creciendo con el tiempo, la ocurrencia del desprendimiento es tanto
más inminente cuanto mayor sea la aceleración del proceso, siguiendo modelos de comportamiento
como los discutidos por diversos autores (por ejemplo, Saito, 1969; Fukuzono, 1985; Voight, 1989).


iv.1.3 Comparación con otras técnicas

Una comparación con otras técnicas y un análisis de alternativas se muestra a continuación: (a) la
principal ventaja en el uso del TLS frente a otras técnicas de adquisición de datos de tipo puntual
(como por ejemplo el GPS diferencial, extensómetros, estación total, etc.), es el amplio
recubrimiento y resolución obtenible con el primero de ellos; (b) otros sensores como el LIDAR
aéreo no permite obtener una elevada densidad de información en zonas escarpadas, siendo estas
zonas en donde suelen producirse los desprendimientos; (c) comparado con la fotogrametría terrestre,
el TLS permite obtener DEMs de un modo más preciso y más rápido (ver, por ejemplo Bitelli et al.,
2004). La combinación del TLS y/o fotogrametría terrestre con un sensor aéreo (ver por ejemplo
Ruiz et al., 2004; Sturzenegger et al., 2007) permite solucionar los problemas de oclusión o zonas de
sombra; (d) una alternativa al empleo del TLS para la detección de desplazamientos precursores



Extended summary in Spanish




Página xviii Tesis doctoral
podría ser el Ground-Based Syuthetic Aperture Radar (GB-SAR). Esta tecnología proporciona
medidas mas precisas que las obtenidas por el TLS tras aplicar el promedio NN descrito en el
capítulo 5 (en laderas naturales se ha obtenido una precisión milimétrica (Noferini et al., 2007)
frente a centimétrica en el caso del TLS). Sin embargo, la resolución espacial obtenible con un GB-
RADAR suele oscilar entre un orden métrico / decamétrico (Tarchi et al., 2005; Corsini et al., 2006;
Teza et al., 2008), siendo esta una importante limitación para la detección temprana de
desprendimientos de rocas de mayor frecuencia y menor magnitud, como los estudiados en esta
investigación.



iv.2 Detección de caídas de rocas
El TLS permite obtener de una manera muy precisa la magnitud y frecuencia de los
desprendimientos durante un determinado periodo de estudio (Rosser et al., 2005; Lim et al., 2006).
Sin embargo, la extrapolación de estos resultados al registro completo no es posible debido a que el
indicador medido (es decir, los desprendimientos) no se producen de manera continua ni en el
espacio ni en el tiempo. Este sesgo poblacional es tanto mayor cuanto más reducido es el periodo de
tiempo considerado y cuanto más reducida es la frecuencia de desprendimientos.

Las diferentes litologías, familias de discontinuidades y escenario geomorfológico controlaron el
distinto comportamiento encontrado en los desprendimientos de Castellfollit de la Roca y Puigcercós,
especialmente en lo relativo a frecuencia y morfología de los desprendimientos: (a) la frecuencia de
desprendimientos fue considerablemente inferior en Castellfollit de la Roca que en Puigcercós, por
lo que para intentar obtener una mejor estimación de la relación magnitud vs. frecuencia
característica de los desprendimientos que se producen en esta zona se optó por una combinación de
los resultados obtenidos del monitoreo con el TLS junto a aquellos obtenidos mediante el registro
histórico; (b) la morfología de los desprendimientos se encuentra condicionada por la litología: por
un lado, los desprendimientos en materiales frágiles (basaltos y calcarenitas) fueron delimitados por
la orientación de los planos de discontinuidad. Por otro lado, los materiales dúctiles (margas)
presentan una erosión mas continua en el tiempo, encontrándose los límites del desprendimiento
menos definidos.

A pesar de la relación encontrada en Puigcercós entre la lluvia promedio y la ocurrencia de
desprendimientos (ver sección 6.3.1), no se obtuvieron datos suficientes para poder establecer un
umbral a partir del cual se puedan producir desprendimientos. El desprendimiento de mayor
magnitud registrado en Puigcercós fue de 87 m
3
. La máxima intensidad de lluvia que pudo
desencadenar este desprendimiento fue alcanzada en otras tres ocasiones sin que se produjera ningún
desprendimiento mayor de 5m
3
. El mismo patrón fue observado en Castellfollit de la Roca. Estos
resultados son consistentes con aquellos obtenidos por Lim et al. (2009), en donde se demuestra que
el factor de correlación entre la ocurrencia de desprendimientos con diversas variables
meteorológicas (intensidad de lluvia, lluvia promedio, temperatura, etc.) es tanto menor cuanto



Appendix 5




Antonio Abellán Fernández Página xix
mayor es la magnitud del desprendimiento. Un factor preparatorio parece ser , por tanto, necesario
para la ocurrencia de los eventos de mayores dimensiones, como por ejemplo la existencia de una
deformación precursora analizada en esta investigación.



iv.3 Deformación precursora
El ensayo experimental mostrado en el capitulo 5 permitió desarrollar metodologías para discriminar
el “ruido” de la medida en si misma. La sensibilidad de los resultados al número de puntos
promediados (k) fue estudiada: mientras que valores reducidos de k no permitían filtrar
suficientemente el ruido, los valores elevados de este parámetro enmascaraban desplazamientos
locales. Finalmente se obtuvo un valor de compromiso entre resolución espacial y filtrado del ruido
correspondiente a un valor de k=24 (5x5 NN).

Tal y como se ha comentado en otras partes de este resumen, la detección de pequeñas
deformaciones se ha identificado como indicador precursor a la ocurrencia de un desprendimiento.
Es importante mencionar que se ha detectado un número considerablemente superior de zonas con
deformación precursora en Puigcercós que en Castellfollit. Esto puede ser debido alguna de las
siguientes razones: (a) el comportamiento dúctil de los materiales de Puigcercós puede favorecer una
mayor deformación precursora (deformación plástica); (b) puesto que la frecuencia de
desprendimientos ha sido mucho mayor en Puigcercós, también lo es la frecuencia de zonas en la
que se ha detectado la deformación precursora.

La aplicación de esta metodología al evento de 50m3 en Castellfollit de la Roca (Abril 2007,
capítulo 5) remarcó la utilidad de este método para la caracterización de pequeñas deformaciones en
casos reales. Sin embargo, la precisión del método es considerablemente inferior en laderas naturales
que en las condiciones “óptimas” simuladas en el capítulo 5 (6.4 frente a 1.3 mm, respectivamente).
Esto puede ser debido a una mayor distancia, reflectividad del material, morfología de la superficie
estudiada y ángulo de incidencia (Petrie and Toth, 2008; Voegtle et al., 2008; Abellán et al., 2009;
Lichti, 2007, respectivamente).

La aplicación de esta técnica en Puigcercós permitió la detección de cuatro zonas con la existencia
de deformación precursora creciente (ver resumen de los resultados en la sección iii.4). La
metodología explicada en la Publicación D (capítulo 6) para la detección de zonas con deformación
precursora creciente y predicción de nuevos desprendimientos, está teniendo continuidad en la
actualidad. Nuevas campañas de campo realizadas durante Febrero y Marzo del 2009 (ver Rodríguez
et al., accepted) nos permitieron detectar otras áreas con deformación precursora creciente en el
tiempo (ver tabla iv).







Extended summary in Spanish




Página xx Tesis doctoral

Deformación precursora Nombre Localidad de
estudio
Volumen
(m
3
) Extensión
máx. (cm.)
Periodo (días)
Ocurrencia
desprendimi
ento
Predicción /
Análisis
retrospectivo
Evento I
Puigcercós 0.8 4 P
iv
(203) SI
Análisis
retrospectivo
Área J
Puigcercós - 3 P
iv a v
(250) - Predicción
Área K
Puigcercós - 3 P
iv a v
(250) - Predicción
Tabla iv. Áreas con deformación precursora creciente detectadas en campañas de campo posteriores a la
publicación D. Estas nuevas áreas confirman la posibilidad de detectar la deformación previa a la ocurrencia de
desprendimientos.



iv.4 Predicción de caídas de rocas.
Respecto a la predicción espacial basada en la detección de una deformación precursora, es
importante recalcar los siguientes aspectos: (a) tal y como queda recogido en secciones previas, se
han detectado varios desprendimientos de magnitudes comprendidas entre 0’1 y 5 m
3
sin que se haya
observado una deformación precursora. Este resultado implica que, o bien la deformación no ha
podido ser detectada de antemano (ver figura 6.4) o bien que sencillamente no ha existido, tal y
como se comenta en la sección 6.4; (b) el desplazamiento precursor detectado en Puigcercós se ha
correspondido a un mecanismo de rotura de tipo toppling. La deformación precursora característica
de otros mecanismos de rotura (fall, slide) podrían ser más difícilmente detectables; (c) mientras que
dos de las tres áreas con deformación creciente seguidas de la ocurrencia de un desprendimiento
fueron detectadas mediante un análisis retrospectivo (desprendimiento de Abril del 2007 en
Castellfollit y Área I en Puigcercós), únicamente fue predicho de antemano la ocurrencia de un
desprendimiento en el área de Puigcercós (área B); sin embargo, (d) la ocurrencia de un
desprendimiento es altamente esperable en las cinco zonas de Puigcercós en las que se está
monitorizando la deformación precursora en la actualidad (áreas F, G, H, J, K).

Por otro lado, se ha intentado conectar la extensión del desplazamiento precursor con la magnitud
del volumen movilizado. Se ha observado que existe una correlación directa de tipo exponencial
entre ambos parámetros. Estas observaciones son consistentes con otros desprendimientos con el
mismo mecanismo de rotura (toppling) pero de dimensiones considerablemente superiores (de 10
3
a
10
7
m
3
). Estos ejemplos se encuentran en (Vibert et al., 1988; CETE, 1990; Glawe et al., 1991;
Glawe and Lotter, 1996; Lotter et al., 1998; Zvelebil and Moser, 2001; Rose and Hungr, 2007). Del
mismo modo, estos resultados son consistentes con los que se discuten en Maruyama and Kozima
(1994) al respecto de deslizamientos de gran tamaño en materiales dúctiles (tipo suelo).

La aplicación del TLS para la predicción de desprendimientos mediante la detección de la
deformación precursora debe continuar validándose mediante: (a) la continuación del estudio para
las 5 áreas detectadas en Puigcercós con deformación precursora creciente en el tiempo (áreas F, G,
H, J, K); (b) la detección de nuevos casos de estudio para distintas litologías, y (c) la detección de



Appendix 5




Antonio Abellán Fernández Página xxi
nuevos casos de estudio para distintos mecanismos de rotura. Estos resultados, son de gran
importancia para la predicción de desprendimientos sobre grandes extensiones del terreno mediante
la futura implementación de sistemas de alerta temprana.












v. CONCLUSIONES


APLICACIONES DEL TLS

La utilización del novedoso instrumental Láser Scanner Terrestre (TLS) está permitiendo en
actualidad una mejora considerable en el estudio y comprensión del fenómeno de caídas de
rocas. Las principales aplicaciones del TLS en laderas rocosas discutidas en este y otros
trabajos son: (a) la obtención de un modelo 3D de alta resolución y precisión en zonas
escarpadas; (b) la simulación en 3D de las trayectorias de caídas de rocas; (c) la obtención
de la orientación de los planos de discontinuidad de la ladera; (d) el monitoreo de
movimientos de ladera; (e) la obtención de la geometría, localización espacial, volumen y
frecuencia de desprendimientos; (f) la detección de un indicador precursor a la a la
ocurrencia de un desprendimiento; (g) la predicción de la localización espacial de futuros
desprendimientos.


DETECCIÓN DE DESPRENDIMIENTOS

La adquisición secuencial de nubes de puntos ha permitido localizar con precisión aquellas
áreas de la ladera en las que se han producido desprendimientos, así como estudiar sus
magnitudes y frecuencias.

En Castellfollit de la Roca se han detectado: seis desprendimientos de columnas
basálticas aisladas con volúmenes de cada una de ellas inferior a 1.5 m
3
y un
desprendimiento de un grupo de columnas con un volumen de 50 m
3
.




Extended summary in Spanish




Página xxii Tesis doctoral
En Puigcercós se ha detectado una frecuencia de desprendimientos considerablemente
superior a la zona anterior: se han detectado un total de 42 desprendimientos con
volúmenes comprendidos entre 10
-3
y 10
2
m
3
. Se ha observado una dependencia de la
litología en la morfología y magnitud de los desprendimientos. Se ha observado una
dependencia de la lluvia promedio en la frecuencia de los desprendimientos.


DETECCIÓN DE PEQUEÑOS DESPLAZAMIENTOS

Se ha demostrado la capacidad del TLS para detectar deformaciones de orden
subcentimétrico en un escenario artificial. La precisión del método ha sido
considerablemente aumentada mediante un filtrado del error instrumental basado en un
promedio de los valores de los puntos vecinos (σ
NN
: 1.3 mm. vs. σ
RAW
: 7.2 mm; distancia:
50 metros).

Se ha demostrado la posibilidad de utilización del TLS para detectar deformaciones de
orden centimétrico en laderas naturales. Esta detección se ha realizado en ocho áreas: un
área en el acantilado de Castellfollit de la Roca y siete áreas en el escarpe de Puigcercós.

La detección de dichas deformaciones ha sido empleada como indicador precursor de caídas
de rocas (slope creep, Terzaghi, 1950).



PREDICCIÓN DE DESPRENDIMIENTOS

De las ocho áreas en las que se han detectado deformaciones de pequeñas dimensiones: (a)
se ha producido un desprendimiento en tres de ellas (B, i, I); (b) las cinco zonas restantes (F,
G, H, J, K) muestran un comportamiento similar a las zonas anteriores, por lo que
ocurrencia de un desprendimiento es altamente esperable. Además, se han observado
desprendimientos de dimensiones decimétricas en la parte superior de dichas zonas, siendo
estos resultados compatibles con los indicadores precursores discutidos en Rosser et al.,
(2007).

La detección de dicho indicador precursor no ha sido posible en todos los desprendimientos
ocurridos durante el periodo de estudio. Esto puede ser debido a una frecuencia de
adquisición de datos insuficiente o a que, en algunos casos, el valor de la deformación
precursora sea inferior a la precisión instrumental.

La deformación precursora se ve influenciada por el mecanismo de rotura, litología y
volumen movilizado: los valores más elevados de duración y dimensiones de la
deformación precursora se corresponden con mecanismos de rotura tipo toppling, materiales
dúctiles y volúmenes considerables. Estos resultados son consistentes con los observados en
desprendimientos de mayores dimensiones y menor frecuencia (e.g. Zvelebil and Moser.,
2001; Crosta and Agliardi, 2004; Hungr et al., 2007).



Appendix 5




Antonio Abellán Fernández Página xxiii
IMPLICACIONES DE LA INVESTIGACIÓN

Las implicaciones de este estudio son de gran importancia para la detección temprana de
desprendimientos: la predicción de la localización espacial puede combinarse con una
predicción temporal en base a leyes de comportamiento definidas por diversos autores
(Saito, 1969; Fukuzono, 1985; Voight, 1989). La implementación de esta metodología en un
sistema de alerta temprana (rockfall early warning system) permitiría conseguir una mejora
sustancial en la gestión del riesgo debido a desprendimientos.




































Extended summary in Spanish




Página xxiv Tesis doctoral






























Improvementsinourunderstandingofrockfall
phenomenonbyTerrestrialLaserScanning

Emphasisonchangedetectionand
itsapplicationtospatialprediction

AntonioAbellán

Supervisor:
Dr.JoanManuelVilaplana

Barcelona,July2009