Monthly Archives: September 2013

Sarah Derendinger: Etude hydrochimique et modélisation géologique de l’aquifère et du tassement de terrain de la Chaux dans le vallon de Nant (VD).

Sarah Derendinger
Co-supervisors: Prof. Michel Jaboyedoff, Prof. Hans-Rudolf Pfeifer
Expert : Prof. Torsten Vennemann

Ce travail porte sur l’analyse hydrochimique et la modélisation géologique de l’aquifère et du tassement de terrain de la Chaux dans le vallon de Nant (VD).

Une campagne de terrain a été réalisée entre avril et septembre 2007, afin de récolter des échantillons d’eau destinés aux analyses des ions en solution par chromatographie ionique et aux analyses d’isotopes stables de l’oxygène et de l’hydrogène. Deux sondages par électromagnétisme transitoire ont également été effectués en vue de déterminer la profondeur du niveau de base du tassement de terrain grâce à la méthode du Sloping Local Base Level (SLBL).

Les résultats des analyses chromatographiques montrent que la composition en ions dissouts des eaux en des différentes sources échantillonnées est bien corrélée à la lithologie de la zone. Leur origine est donc très locale et il n’y a pas d’apports extérieurs. Par contre, une zonation entre les parties amont et aval du vallon est visible. En effet, la partie amont présente des teneurs plus importantes en magnésium et en sulfates, provenant de la dissolution de roches et minéraux magnésiens, ainsi que de la dissolution de pyrite, probablement contenus dans les Flysch parautochtones constituant le substratum de la nappe de Morcles. Les analyses des isotopes stables permettent d’établir un temps de rétention des eaux dans l’aquifère d’environ 4 mois, ce qui est confirmé par l’analyse de la variation temporelle au niveau mensuel des concentrations en ions dissouts dans les eaux de sources.

La modélisation géologique a permis d’établir la profondeur maximale de l’aquifère à environ 100 m, et celles du tassement de terrain et du remplissage du fond du vallon aux alentours de 200 m. Les volumes établis par la suite sont les suivants : l’aquifère mesure 47’508’325 m3, le tassement de terrain 165’694’075 m3 et le remplissage du fond du vallon, dans la zone analysée, 163’552’250 m3.

Tiziana Egli: Cartes de danger de chutes de blocs sur le village des Morgnes (Bagnes)

Tiziana Egli
Co-supervisors : Dr. Vincent Labiouse, Prof. Michel Jaboyedoff
Experts : Dr. François Joseph Baillifard, MSc. Clément Michoud

The alpine village of Les Morgnes is dominated by a long cliff. Some previous reports and blocks present in the forest below the cliff informs that the village is threatened by this cliff. The aim of this study is to establish the rockfall hazard map for this place combining two different methodologies: the Matterock approach and the CADANAV approach.

The studied cliff is located in an old glacial valley and it’s surrounded by landslides. The cliff is mainly composed by quartzite and past tectonic conditions has created a significant schistosity. According to the Matterock methodology, field campaigns and LiDAR scans establish the structural conditions and the potential localized sources. The susceptibility mapping was performed with a software according to kinematic tests. Toppling and wedge sliding are the principal failure mechanisms. 2D trajectography were simulated along predefined paths. The hazard assessment, based on the trajectography studies, demonstrate that no building is in a dangerous zone. According to the CADANAV approach, 3D trajectography studies were performed. The semi-automatic hazard assessment shows that depending on the return period, some buildings are located in a high danger zone.

Maria Ponzio: Etude des instabilités du Val Canaria (Tessin, Suisse) et analyse de la stabilité de l’éboulement de Sotto Corte

Maria Ponzio
Supervisor : Prof. Michel Jaboyedoff
Rapporteur : Prof. Stefan Schmalholz

La zone d’étude se trouve dans le Val Canaria, au Nord du Tessin (proche du village d’Airolo). Cette zone est caractérisée par plusieurs phénomènes d’instabilités.

Le but de ce travail de master est de comprendre les mécanismes qui agissent sur la formation des instabilités. Pour l’atteindre une étude basée sur les observations de terrain et les méthodes de télédétection (Lidar aérien et Lidar terrestre) a été réalisé. Une carte des principales instabilités est proposée. Les deux flancs de la vallée sont affectés par une Déformation Gravitaire Profonde de Versant (DGPV) qui comprend, en son sein, différentes autres instabilités, notamment des glissements profonds, des glissements superficiels, d’éboulements rocheux et des érosions de berge.

Une carte géologique a été aussi été réalisée pour comprendre la potentielle corrélation entre instabilités et lithologies. De plus, cette carte a été utilisée pour créer un modèle géologique en 3D. Les principales lithologies présentes dans la zone d’études sont des roches sédimentaires mésozoïques (dolomie, cornieule, gypse, anhydrite, Bündnerschiefer et Quartenschiefer), ainsi que les roches constituent le massif du Gothard (métapélites, gneiss, amphibolites) et la nappe du Lukmanier (gneiss et micaschistes).

Une dynamique de l’activité instable du flanc droit voit la présence de plusieurs plans de glissements qui sont liés au contact entre les dépôts non consolidés quaternaires (principalement des colluvial) et la roche, la dissolution des roches carbonatées (dolomie et cornieule) et des roches sulfatées (gypse, anhydrite) et l’activité tectonique de compression. Les facteurs dégradants agissant sur le versant sont principalement les infiltrations d’eau, l’état de la fracturation de la roche et l’érosion de pied du versant par la rivière.

Le flanc gauche se caractérise par une zone affectée par un glissement superficiel très étendu qui est principalement causé par l’activité érosive de la rivière. Les plans de glissements sont multiples et le glissement est en train d’évoluer en direction de l’amont de la pente. Comme dans le cas du flanc droit, les facteurs dégradants sont principalement la présence d’eau dans la couche de matériel meuble ainsi que la forte fracturation de la roche.

L’étude en détail d’un éboulement rocheux de Sotto Corte dans le flanc gauche a permis d’interpréter le mécanisme de rupture, ce qui permet d’émettre des hypothèses sur une possible évolution future de cette instabilité.

Le monitorage par LiDAR terrestre a montré que l’activité de chutes de blocs est importante dans la partie supérieure de la niche d’arrachement. Dans le cône d’éboulis, les ruissellements d’eau superficielle engendrent l’érosion du matériel fin déposé lors de l’éboulement survenu en 2009. Les discontinuités présentes dans la zone peuvent géométriquement provoquer des glissements de type plan ou dièdre. La stratigraphie plonge dans la même direction de la pente.

Le modèle géomécanique, par la méthode des éléments finis, a permis de comprendre la dynamique qui a amené à la rupture du 2009. Le plan de cisaillement se trouvera à la limite entre le gypse et l’anhydrite et les fractures, ne contrôlent que partiellement l’instabilité. Ces informations ont permis permettre d’estimer les volumes instables encore présents dans la zone. Deux compartiments avec des volumes maximaux de 90’000 m3 et 26’000 m3 ont été identifiés.

Aucun facteur déclenchant n’a été identifié pour cette instabilité. Plusieurs facteurs déstabilisant sont présents, en particulier l’effet de l’eau, qui joue un rôle principal pour l’instabilité. La circulation dans le massif provoque la dissolution du gypse, ainsi que l’hydratation de l’anhydrite, qui se transforme en gypse. L’érosion du pied de la pente par la rivière amène à une perte de stabilité. Le phénomène de gel-dégel provoque la fracturation de la roche, et en conséquence, augmente sa perméabilité. Le modèle géomécanique montre aussi que la dégradation des paramètres de résistance au cisaillement de la roche et la présence de la nappe phréatique provoquent la rupture.

Une évolution possible future est la propagation progressive de cette instabilité en direction le Nord-Est, zone où se trouvent plusieurs failles et fractures. A plus à courte terme trois possibles scénarios sont possibles. Un effondrement simultané de deux compartiments ou l’effondrement individu de l’un au l’autre des compartiments.

Camille Perraudin: Characterization of Montset rock slope instability in Hérémence valley.

Camille Perraudin
Supervisor: Prof. Michel Jayboyedoff
Experts: Dr. Marc-Henri Derron, MSc. Dario Carrea

The study focuses on the characterization of Montset rock slope instability in Hérémence valley. The first part of the work aims to determine, on the basis of geomorphological and geometrical characteristics, what kind of instability is that of Montset, which at first looks like a Deep Seated Gravitational Slope Deformation. The next step is to determine its level of activity, i.e. if the movement can be measured and / or observed, and, if necessary, to locate them. Signs of activity, in terms of surface and deep instabilities are detected by field observations and also from the information obtained with LiDAR. This provides a series of measurements of high density accuracy for the creation of a point cloud of the topography. The detected instabilities are described and their susceptibility to failure is evaluated. A hazard map for rockfall in the study area has been created. Following this, a model of instability is constructed. This model incorporates field observations, displacement measurements, information from the geological map and subsurface data. These data were acquired during the visit of a hydraulic gallery underlying the eastern part of the Montset. Finally, based on the synthesis of information from the study, assumptions about the history of instability and the reasons that led to its development are proposed.

Line Rouyet: Monitoring and characterization of rock slope instabilities in Norway using GB-InSAR.

Line Rouyet
Co-supervisors: Prof. Michel Jayboyedoff, Dr. Marc-Henri Derron
Experts: Dr. Lene Kristensen, MSc. Clément Michoud

The master thesis named “Monitoring and characterization of rock slope instabilities in Norway using GBInSAR (Ground-Based Interferometric Synthetic Aperture Radar)” aims (1) to improve the knowledge about the advantages and limitations of a GB-InSAR system (©LiSALab-Ellegi) for the monitoring of instable rock slopes; (2) to contribute to the understanding of two large rock instabilities, especially in terms of behavior and geometry.

The work includes (1) the processing and interpretation of GB-InSAR data for two main case studies of major rockslides in Norway, as well as for a third test case involving rockfalls; (2) the comparison of GB-InSAR results with other available monitoring data (GPS, crack/extenso-meters, lasers-reflectors, tiltmeters and meteorological data), as well as the integration of structural analyzes (using Terrestrial and Aerial Laser Scanning data); (3) the global synthesis of the results and the comparison with the conclusions of precedent studies in order to reach the two objectives previously highlighted.

The first case study is the site of Indre Nordnes (Troms County, northern Norway) which is affected by a large rockslide above Lyngenfjord generating a significant risk of tsunami. Its volume is estimated to be around 10-15 millions of m3. Two intermittent GB-InSAR campaigns were performed in summer 2011 and their results are compared and integrated with the large monitoring database available since the end of 2007 for the crackmeters, the beginning of 2010 for the GPS, since mid-2011 for the extensometers and tiltmeters.

Generally, the results show that the rockslide is slow (6-7 mm/year) with accelerations during the thaw periods (May-June). The year 2011 is an exception with an important acceleration in June (up to 14 mm in only 2 weeks according to GB-InSAR data). These values can be related to the exceptionally high temperatures in 2011. Moreover, GB-InSAR data allow giving new information about the extent and the repartition of movements of instable area. Finally, thanks to the structural analyzes two discontinuities sets are supposed to be mainly involved in the sliding. It suggests a stair-steps geometry of the sliding surface.

The second case study is the large Børa/Mannen area (Møre of Romsdal County, western Norway), which includes the rockslide of Mannen (15-25 millions of m3) as well as the Børa plateau located on the southwest side of Mannen. Because Børa area showed some signs of activity, intermittent GB-InSAR campaigns were performed in summer 2011 and 2012. Their results are compared and integrated with the large monitoring database including GPS campaigns between 2003 and 2010 on Børa plateau, various in situ data in Mannen since the end of 2010 and permanent GB-InSAR data since the beginning of 2010. This second GB-InSAR system is located about 2 km more northeastern than the one of Børa and aims to monitor the Mannen rockslide. It has the advantage to be in overlap with the Børa GB-InSAR view and to provide continuous images between 2010 and 2013.

Overall, a good concordance is found between the results of the two GB-InSAR systems. In the overlap part, a coherent and progressive pattern is highlighted in the middle of the slope. During the August 2011 campaign, positive displacements (increasing of distance between the GB-InSAR and the target) are recorded and exceed +8 mm in two weeks. With data integration at long-term, it is shown that this part is affected by large variations and inversions of movements, which can be related to seasons. These effects could be linked to water circulation in the rock slope, but the explanations are not yet clear.

For the Mannen rockslide, it is possible to distinguish two parts with different behaviors in terms of velocities and orientations of movement. It highlights that the most upper and southwestern part is the most active (>20 mm/year). Otherwise, thanks to the structural analyzes three discontinuities sets are supposed to be mainly involved in the sliding.

For Børa area, GB-InSAR recorded higher displacements (25-90 mm/year) than those estimated from GPS campaigns (mainly no significant displacements with a maximum velocity of 14.6 mm/year). In any case, it appears that the area is generally moving slowly and quite regularly.

Finally, high variations at short-term and long-term are visible on the interferograms which cast doubts on the data quality. A case of loss of coherence probably due to snowfalls in October is also detected. This underlines the high potential of researches to improve the removing of atmospheric effects and the capability of the GBInSAR in case of external perturbations.

The third test case of Sunndalsøra-Oppdølsstranda (site C along the road 70 between the two localities, Møre of Romsdal County, western Norway) is studied in order to improve the knowledge of the potential of GBInSAR for the detection and analysis of rockfalls. Between April 2010 and October 2011, five GB-InSAR campaigns were performed and give information about a rockfall event of 15-20 m3 which occurred the 8th of June 2011. The results confirm the capability of the system to record precursor movements to failure, but highlighted also various operational limitations.

Overall, the joined analysis of the three different case studies allows having a good overview of the potential and limitations of the GB-InSAR system. Moreover, because of their respective advantages and limitations (in terms of kind of recorded information, temporal sampling, accuracy, spatial coverage, etc.), it appears that the GB-InSAR and the in situ monitoring devices have a good complementarity. The integration of their results therefore allows giving new inputs about the behavior and geometry of the Indre Nordnes and Børa/Mannen sites.

Léonard Jaton: Development and application of a hydrological and slope stability model using Matlab

Léonard Jaton
Supervisor : Prof. Michel Jaboyedoff
Expert : Pascal Horton

This master thesis is focused on the development and application of a hydrological and slope stability model for use on catchment basins.

In the field of natural risks analysis, and particularly those of quick rises in river water levels and resulting floodings as well as landslides, a modeling of the hydraulic processes taking place in a catchment area during periods of rainfalls is a crucial point. This allows a better comprehension of the way water is transported downhill, and so permits to better identify and predict dangerous areas and events.

A computer program is therefore developed in Matlab to model the various intervening principles and obtain results of flow rates and slope stability from precipitation data on the area of study.

The following paper is made up of a first chapter presenting the theoretical concepts used as well as their adaptation for the model. The program and its functions are then detailed and explained, fol-lowed by a section of tests and application of the model to firstly theoretical then natural basins. In a last chapter, several criticisms of the model and other development ideas are proposed.

Jérémie Voumard: Simulation dynamique du trafic routier pour l’estimation du risque sur les routes de montagne

Jérémie Voumard
Supervisor : Prof. Michel Jaboyedoff
Experts : Dr. Marc-Henri Derron, Dr. Luuk Dorren, Olivier Caspar

The main theme of this Master thesis is the calculation of risk generated by natural hazards on roads. While it is common to assess risk incurred by the road users through static calculations, a new approach based on a dynamic calculation of risk is proposed here. Risk is usually calculated at a so-called macroscopic scale, i.e. with general parameters representing the road and its related traffic. The new approach presented here is to consider the traffic at a microscopic scale, taking into account risk for each vehicle traveling on the road. This new dynamic approach ensures a more realistic application of the theoretical concept of risk calculation. For this purpose, a road traffic model was created and implemented into a simulator specially conceived in this Master thesis. An analysis of various models used for traffic simulation, a reminder of fundamental concepts of computer simulation and a presentation of software able to simulate road traffic are also presented.

In a first, the existing context around risk calculation on the roads is explained and highlighted by:

  • A reminder about the history of roads with a focus on mountain roads in Switzerland,
  • A description of natural hazards that may occured in the three geographical areas of Switzerland and worldwide,
  • An inventory of natural events on roads in Switzerland between winter 2011 and summer 2012,
  • An analysis of the Swiss inventory of road accidents caused by rockfall.

The second part is about the numerical simulator itself. The simulator, developed using the programming software MATLAB®, is based on two different scenarios: the first simulating the effect of a rock on the road traffic and the second modeling an alternating traffic governed by traffic lights. The simulator includes two categories of vehicles for which the dynamic variables and those related to traffic and speed limits are freely configurable. Reductions in visibility and curve speed are taken into account.

Dynamic risk is based on the vehicles number in the danger zone as well as their speed. Standard risk is also calculated to compare the two results. Thereby, an almost infinite number of real traffic situations can be simulated, such as:

  • A straight road without obstacles,
  • A curve section with a high rate of slow vehicles,
  • A rockfall cutting one or two lanes,
  • A road where traffic is controlled by two alternating traffic lights, etc.

In the third part, the simulator model was calibrated and validated with data from the main road linking the town of Aigle in the Chablais Vaudois to the tourist resort of Les Diablerets. Various data relating to the speed of vehicles and their behavior were collected in the field on several sections relevant to the evaluation of natural hazards threat on the road. Otherwise, a history of the road, completed with traffic statistics and socio-economic data contributed to understand the acceptability of risks on the road by the population.

The results of the dynamic calculation of risk from different simulations were compared with results of standard calculations and the most dangerous situations, for instance a column of vehicles stopped at a red light in a dangerous area, were highlighted.

To conclude, a critical assessment of the simulator, with its perspectives, is proposed with a discussion about the risk calculation on the roads.

Virgile Nanchen: Application de la technologie LIDAR à la méthodologie Matterock

Virgile Nanchen
Co-supervisors : Prof. Michel Jaboyedoff
Experts : Alex Loye, Andrea Pedrazzini

The rockslide of 1991 in Randa motivated the development of Matterock methodology (Rouiller et al. 1998). Based on the principle that there is no hazards without discontinuity, this evaluation method of rock walls is based of the structural pattern of the cliff to identify the compartments favourable to rockslide Within the framework of the methodology and as a complement to field studies, some altitude numerical models and statistical methods has been used in order to create automatically hazard’s map.

Since the establishment of the methodology, tools for automatic detection have been highly improved, the resolution of terrain numerical models has increased, and new instruments based on laser technology have been developed, enabling to make 3D models of rock faces. The LIDAR (Light Detection and Ranging) system has shed a new light in terrain analysis. It enables to create point’s cloud with a centimetre resolution, representing the surface geometry of rock walls.

Jointly to the development of such a technology, several methods, based on clusters of points and high resolution MNTs, have been developed in order to identify the structural pattern of the cliff. Within the framework of this work, several on these methods have been applied on the site where the Matterock methodology has been developed. Results are compared to those of 1998 in order to outline the contribution of these new technologies. The objective of this work is to evaluate if such new tools are comparable to the Matterock methodology.

Romain Chenillot: Analyse des aléas rocheux des falaises du Grand Bois du Ban, Lavey Village

Romain Chenillot
Co-supervisors: Prof. Michel Jaboyedoff, Prof. Vincent Labiouse, Dr. Mario Sartori

This analysis of rock falls hazards relates to cliffs of the Grand Bois du Ban, located by the city of Lavey, in the canton of Vaud, Switzerland. The activity of this site was established on the basis of observations of blocks and traces of impacts carried out in wood. The purpose of this study is to determine the various elements characteristic of the phenomenon, thus allowing the development of hazard maps based on two softwares of trajectographic modelings and two methods of zoning.

The cliffs are located in the Helvetic’s tectonic whole, on the level of the Morcles’ klippe. They are made primarily of massive and compact limestones. It was made possible to characterize five families of discontinuities, and the stratigraphy based on the structural analysis, the field’s observations and the LiDAR scans treated by Coltop3D. The whole zone is affected by a rather important fracturing, but also by the presence of folds.

Field’s observations and LiDAR analysis made it possible to determine three formations processes of dihedral slip and a process of planar slip related to stratigraphy. Matterocking software could give potential rock instabilities on the basis of these observations.

At this point, the trajectographic’s tests is feasible. Two softwares were used. Initially Rocfall, limited to a modeling 2D, then Rockyfor3D allowing the taking into account of three-dimensional trajectories. The comparison of the results between these two models shows good similarities of the blocks stopping limits. Although the disparities of the energy intensities are sometimes important.

Hazard maps were carried out according to two methodologies respectively the Matterock Modified based on the results of Rocfall and the new Cadanav methodology based on Rockyfor3D. The analysis of the results shows a less favourable mapping for Modified Matterock, especially if blocks of great volumes are taken into account with one great period of return. This element brings the fore on one of the essential differences between the two methodologies. Namely taking into direct account the period of return in the case of Cadanav, which is left aside using Modified Matterock.

Lastly a second series of hazard maps was carried out in order to test the impact of the variation of a parameter on the results of mappings using the Original Matterock which takes all blocks into account. This differs from the Modified Matterock withdraws the 10 % of the modeled extreme blocks. Concerning Cadanav, the periods of return values estimated based on field’s observations were divided by 5. The results show important mapping variations, illustrating these methods great sensitivity to the fluctuation of its parameters mainly based on estimations.

As a conclusion of this analysis of rock falls the village of Lavey is located out the dangerous perimeter. However all limits previously mentioned must be taken into account and decisions made accordingly.

Tristan Burnand: Étude des processus érosifs et bilan sédimentaire au Vallon de Nant

Tristan Burnand
Supervisor: Prof. Michel Jaboyedoff
Expert: Alexandre Loye

This master thesis studies the Vallon de Nant, a glacial valley located in the Vaud Alps on the border with the canton of Valais. Within this valley, the different erosion processes are studied.

In context, the processes related to erosion are mapped on the basis of field visits, orthophotos and digital terrain model (DTM) high precision (spatial resolution of 1 m).

A qualitative study, by means of indicators such as the stream power and the erosion index are used globally in the Vallon. This allows you to split the catchment into three distinct zones (A, B and C), governed by different dominant processes. Within each zone, a secondary watershed is studied in more detail to test the assumptions for each predetermined area. The first area is dominated by gravity processes and mobilization of sediments. A second area is dominated by large geological structures and high sediment yield. The last zone corresponds to the system of glacial and paraglacial glacier des Martinets.

These three areas have in common, however, that the sediments that eventually come into the riverbed of the Nant Avançon are then transferred down by fluvial processes. The interaction of these erosion processes can be considered as a sediment cascade, where various sedimentary reservoirs and fluxes are combined.

The method of local base level (SLBL) is used to determine the volume of sediment present in the valley floor. This amount is used to calculate an average denudation rate since the glaciers last retreated, and a sediment yield.

A sediment budget, based on the mapping of phenomena established, is performed using the loss of potential energy concept. It is made for a single secondary catchment. The results give an average erosion rate of 0.74 mm/year, with a minimum of 0.24 mm/year and a maximum of 1.24 mm/year.

On the same secondary catchment, modeling the spread of rock falls is made. Then, based on the concept of the sediment cascade modeling the spread of sediment remobilization, by debris flow processes, is undertaken.