Monthly Archives: September 2013

Benjamin Rudaz: Evolution Géomorphodynamique d’un bassin versant torrentiel

Benjamin Rudaz
Supervisor: Prof. Michel Jaboyedoff
Expert: Alexandre Loye

This master thesis is focused on the Saint-Barthélémy stream, in the canton of Valais, Switzerland. It drains a catchment cut in three different geological units forming the Dents du Midi massif. Its massive torrential fan fills the whole width of the Rhône Valley.

Historically, the Saint-Barthélémy has on numerous occasions affected the human activities in the region, by cutting communications, damming the Rhône river, or simply by overflowing its banks on the debris cone. Thus, this work starts with a historical reconstitution of its past activities, with the goal of finding a pattern. The best documented crisis happened between 1926 and 1930. The different crisis recorded are mainly caused by rock avalanches in the upper part of the catchment. They result in a huge sediment supply, in a steep and snow-filled gorge. Debris flows initiate there, amplify in volume in the middle-part of the catchment, and reach the alluvial fan in multiple waves. To stabilize the terrain, and thus avoid future debris flows, the 1930 crisis was followed by an intense correction campaign. The different dams are hereby described, and their current status is evaluated.

The steady-state regime theory, applied on longitudinal profiles, states that for a stream to be in such state, its long profile must be upward concave, and follow a mathematical form, be it exponential, logarithmic or that of a power law. The streams present in the catchment are evaluated on this basis, using theorical longitudinal profile, of the exponential type. Their concavity is calculated, thus giving information about the maturity of said profiles. The longitudinal profiles are compared in their normalized form. The results identify a global pattern of potential sedimentation at the center of the catchment, while the external parts are expected to erode. The desequilibrium is stronger in the northern lateral streams.

A longitudinal profile in steady-state regime should be smooth. Therefore, profile irregularities, or knickpoints, are another sign of unsteady regime. To identify those knickpoints, three methods are used. First, using normalized profiles, the knickpoints are subjectively identified. Two numerical methods are then used to detect brutal change of local slope, relative to the neighbouring segments. The results show more precisely where future erosion is expected. The upper Saint-Barthélémy stream is particularly disturbed. The knickpoints are finally caracterized between geological, geomorphological and human origin.

By orthorectifying aerial photographs dating between 1969 et 1992, the activity of landslides connected to the main stream is measured. Lastly, using a sloping local base level method, the volume trapped behind the check dams is calculated, allowing to estimate mechanical denudation rates for different time frames. These rates (0.14 up to 0.84 mm/year) are in the range of similarly obtained rates in the Swiss Alps.

Anthony Monnier: Etude du danger de chute de blocs le long de la ligne ferroviaire Aigle – Les Diablerets

Anthony Monnier
Supervisor: Prof. Michel Jaboyedoff
Expert: Andrea Pedrazzini

This Master memory focuses on the risk of rockfall along the Aigle – Sépey – Diablerets railway line, in the Swiss Alps.

This memory breaks up into four main parts: Identification of the rockfall sources, risk definition by trajectometric analyses and field observations, the creation of hazard maps, the calculation of the damages cost and the individual risk, protection measures proposals and a calculation cost/effectiveness of these measures.

The most dangerous zones along this railway line have already been protected a few years ago, on the basis of a detailed study and general studies. But it remains three zones classified in 1st priority which were not secured. These three zones are the subject of this memory.

We can find indicative hazard maps in Switzerland, generated by computer based on topographic and geological criteria, which do not have the force of law. They have only an advisory role for the landscape planning.

Securizing the Swiss railway lines against the rockfalls consists in protection measures and monitoring measures in rare cases. There are four principal types of protection against rockfall : the rockfall screen, the rockfall net, the barrage and the gallery.

The hazard identification consists in a consultation of the preceding studies and an investigation of the people directly concerned by the rockfall hazzard, field observations as well as localization of the already fallen blocks and the discontinuities families so being able to determine the mechanisms of rupture. Treatment of the DEM by Matterocking© can confront the field observations with the data-processing modelings. The use of the DEM requires an error evaluation, by confronting it with altimetric measures taken with a differential GPS.

Once the rockfall sources located, it’s time to determine the hazzard, consisting in the propagation and the intensity at each place of the study perimeter. The size of the blocks is estimated according to the fallen blocks and the scanline method. Rotomap3D©, a trajectometry software allows to determine the main rockfall corridors and then extract the profiles which are processing by Rocfall©, a two dimensions trajectometry software.

The results of these simulations make it possible to define the intensity of the danger according to the energy mobilized by rockfalls as well as the propagation and the hazzard of the scenarios which can occur on the three studied zones.

Two types hazard maps are produced: One is based only on the energy mobilized the rockfall simulations, the other uses the same the criterias than the preceding but it includes some conventions and its polygons are smoothed in order to allow a better legibility.

In order to calculate the cost of the potential damage generated by rockfalls in these areas and the individual risk of a person travelling by train, it is necessary to define some scenarios, their consequences, their recurrence time and the localization of their occurrence, according to the characteristics of the rockfalls and topographic criterias.

The protection measures suggested here consist only of rockfall nets. Two alternatives are proposed: a total protection and a protection decreasing to the maximum the risk for a minimum investment. The utility of these alternatives will be evaluated using a calculation cost/effectiveness.

Lastly, a railway line as that studied does not have a much traffic, which makes protection measures less or not profitable. A study way would be the installation of a rockfalls monitoring system. These systems are mainly developed in Canada. The most elaborate of them uses some geophones in order to be able to detect several natural phenomenas as the rockfalls but also exploitation incidents like a derailment or a broken rail.

Fabian Eggertswyler: Analyse de l’évolution d’une commune soumise à des dangers naturels au moyen de techniques numériques. Etude de site: La commune D’ormont-Dessus.

Fabian Eggertswyler
Supervisor: Prof. Michel Jaboyedoff
Experts : Prof. Jean Ruegg, Alexandre Loye, Marj Tonini

Alpine valleys represent a significant living environment and are decisive for economic activities. Construction of buildings and roads in these environments where the soil is not easily workable is increasingly important. However, these regions are highly vulnerable to natural hazards and infrastructure development contributes to increasing the extent of damage caused by natural disasters (OFS & OFEV 2008).

Since 1997, management of natural hazards is part of territory planning. Obviously, the observation of natural hazards in the tenure of the soil affects our security. It is therefore essential to plan the land in order not to aggravate the potential damage in high-risk areas or even, to reduce them on the long range (OFEV & OFS 2007).

Located in the Prealpes, Ormont-Dessus, due to its geomorphology, is subject to various and multiple risks. Mainly by agricultural necessity, these high-risk areas have been long-occupied and exploited (STUCKI E. & ROGNON P. 1998).

The study of this localitie’s evolution affected by natural hazards provides a report indicating how infrastructures and communication channels have been built and developed over the past 60 years. The analysis of their spatial distribution reveals a sharp increase in the number of buildings in the village of Les Diablerets. On the other side, on the slopes Chaussy Para, exposed in the past to many avalanches, the building density has decreased significantly. Furthermore, the proportion of buildings located in areas of potential hazards and mainly affected by landslides, has changed little since bâtiments A deeper investigation for the analysis of territorial development and high-risk.

A deeper investigation for the analysis of territorial development and high-risk areas is possible.

Benoit Fragnol: Le LiDAR et ses applications pour l’étude des chutes de blocs

Benoit Fragnol
Supervisor: Prof. Michel Jaboyedoff

In mountainous areas, rockfall hazards cause many damages to the infrastructures and have a major socio-economic impact. These phenomena are still not well-understood and that is why new technologies are required in order to improve the characterization of the rock mass condition and the failure mechanisms.

Field methods as structural measures, scanline, GSI, JRC are used to characterize geometrically and geomechanically the unstable area but they required a high number of measures to be reliable and on a substantial number of outcrops, which is not always possible.

The LiDAR technology (Light Detection and Ranging), terrestrial and airborne, appears to be a powerful tool for imaging study areas and to work in 3D instead of 2D. It is particularly well-adapted to the analysis of unstable areas because measurements can be taken from a distance up to 1 km to the cliff and limits exposure of the geologists during the field work. It can even overcome the field measurements when the study area is too dangerous or the outcrops cannot be reached. The softwares processing LiDAR data are necessary tools for the identification of the different discontinuity sets. They allow performing a structural analysis of the area based on spatial orientations [Dip direction/Dip] for each point of LiDAR point cloud (Coltop-3D), to give a relative probability of presence of instability depending on geometric relationship between facets of DEM and spatial orientations of discontinuity sets (Matterocking). It is also possible to compare different temporal series of LiDAR data in order to identify rockfalls and to estimate their volumes (Polyworks).

The contribution of LiDAR imaging and data processing are illustrated by the structural analysis performed on the cliffs in front of Huteggen (VS, Suisse) along the main road leading to many ski resorts. This analysis aims to identify potential instabilities and define their failure mechanisms. An estimation of the instable volumes can be made based on LiDAR data combined with field observations. The results of this study allowed writing recommendations for the geological survey of the district in charge of protecting the road against natural hazards.

Florence Delasoie: Études des instabilités de versant de la Barmasse (Valais, Suisse) par une approche pluridisciplinaire.

Florence Delasoie
Co-supervisors: Prof. Michel Jaboyedoff, Dr. François Baillifard
Rapporteur: Prof. Jean-Luc Epard

The DSGSD (Deep Seated Gravitational slope Deformations) phenomena occur in most mountainous landscapes. Despite this phenomen has been recognized worldwide as a major processes of mass wasting, the mechanisms of most mapped DSGSD are not well understood.

In the Bagnes Valley (Valais, Switzerland), 14 DSGSD have been detected. Geologically, the field is situated in metamorphic rocks : quartzite-schists, prasinite. This study focuses on the Barmasse area, which is the seat of a 1.5 square-km-large DSGSD. The deep-seated movements cause shallower landslides with important displacements that produce deformations on the cantonal road connecting the lower Bagnes valley to the Mauvoisin hydroelectric power plant. Moreover, frequent rockfalls are originating from the scarp of these landslides or from their scree slope. Due to the economical importance of the object at risk, the municipality of Bagnes and the canton of Valais set up a monitoring system to monitor the movements of the shallow landslide, and build protection dams to reduce the rockfall risk.

The aim of this study are 1) to characterize the mechanisms of the DSGSD in order to evaluate its activity and its future evolution and 2) to better understand the link between superficial and deep movements.

A multidisciplinary approach is used, containing field mapping, structural analysis, measurements displacement (Extensometers, DGPS and Terrestrial Laser Scanning), dendrochronology, geophysics, hydrogeology,…

The results show that the mechanisms of the DSGSD as well as the shallow landslides is mainly driven by large scale toppling of subvertical sets. 6 families of discontinuities have been detected in the field and in Coltop 3D : J3 (parallel to the foliation S0), J2/J1, J4/J5 (subvertical) and J6 (horizontal). In the main scarp, J3=S0 couple with J4 or J5 makes a wedge mechanism rupture possible. Toppling is the most important mechanism failure along J2, J3 and S0, when the foliation dip toward the South.

The movements measured on the DSGSD itself are low (mm/year), but the displacements of the shallow landslides are very important : they reach about 3-5 m/year in the scree slope. As indicated by dendrochronological analysis, datation on trees located on the area and wood in the borehole, the activity of these landslides is recent (between 2’000 BP and 200 BP).

Postglacial debustering and gravity are the main triggering factiors for the deep-seated movement. The shallow landslides are controlle by rainfalls and snow melting. Today, the blocs deposit creep along the valley and the fracturing continues.

Raja Mastouri: Modelisation 3D des bassins sedimentaires du Sahel Tunisien: Implication Géodynamique et Intérêt Pétrolier

Raja Mastouri
Co-supervisors: Prof. Samir Bouaziz, Prof. Michel Jaboyedoff
Experts: Merj Tonini, Chedly Abbes, Mohamed Moncef Sarbeji

The Tunisian Sahel domain is characterized by geological outcrops that range from Miocene to Quaternary. This area has been considered as a vast plain and as collapsed block. Its morphology shows gradients of up to 200 m, as hills, detailed reliefs, plains and depressions. However, it’s marked by a gently tectonic frame work showing in subsurface more complexity in horst and graben with kilometric scale bounded by systems of branching faults and fold structures.

To better understand the layout of these structures in relation to the sedimentary series, the 3D modeling allows visualizing in three dimensions the geodynamic evolution of Mesozoic and Cenozoic basins of the Tunisian Sahel. It is to build and use a more simplified model closer to reality. This building 3D models and their use remains uncommon in the field of geology and represents a new tool in the study of sedimentary basins.

This study aims to develop a procedural guide that assembles a logical step for achieving 3D geological model surface and subsurface using geographic information systems (GIS) and in particular the software “Arc-GIS” to create a digital model land and a 3D surface model and the software “Surfer” to digitize the basis of subsurface data. This method describes the application of modeling software “GéoShape” that facilitates obtaining a more detailed 3D model of surface and subsurface volume for oil exploration.

Aurélie Pannatier: Méthodologie de détection et d’analyse d’instabilités rocheuses aux échelles régionale et locale. Applications à la vallée de Yosemite (Californie/Etats-Unis) et à l’éperon du Sex Frei (Valais/Suisse).

Aurélie Pannatier
Co-supervisors: Prof. Michel Jaboyedoff, Thierry Oppikofer
Expert: Frank Philippossian
Rapporteur: Prof. François Marillier

Rock instabilities have been studied for a long time, but no strict methodology has been developed to ease their analysis. In this research paper, it has been decided to create two types of methodology related to the scale of the area under study.

The first type of methodology is applied to the regional scale. Its goal is to determine areas that might produce the rockfalls solely using a digital elevation model. It uses the interaction of the following parameters: slope, planar or wedge slidings, and thickness of the erodible mass. A susceptibility map of possible rockfalls is finally obtained focusing field investigations where the susceptibility is the highest.

This methodology has been applied to the Yosemite Valley (CA, USA). This region has been studied for a long time because it is very sensitive to rockfalls. Therefore, it is possible to compare the results of this method with the pre-existing studies in order to judge its coherence and precision.

The second type of methodology is applied to the local scale. It is, in a way, the extension of the regional methodology. After having identified the areas in needs of further investigation, the next method is applied using field observations, digital elevation model or modeling programs. Finally, geomorphologic and structural analysis, a monitoring using the terrestrial LIDAR, a geomechanic analysis, a modeling of the failure mechanisms, and a prospective analysis are performed in order to create a propagation map and an intensity map.

This methodology has been applied to the cliff of the Sex Frei located above the small village of Pralong in the Heremence Valley (Canton of Valais, Switzerland). This cliff has been studied for over thirty years by the “Bureau d’Etude Geologique”. Movements of about 1 cm/year have been detected using a differential GPS.

A study of the entire Heremence Valley indicates that the area under study has a high susceptibility to produce rockfalls. Putting the local methodology into practice showed that the cliff is at limit of stability and that a planar sliding prevails in the mechanism of failure. The propagation and intensity maps created permit to tell with certainty that no isolated block can reach the bottom of the valley, putting the Village of Pralong at risk. However, it is not excluded that a more serious event could occur after an earthquake, then implying a rock avalanche that could possibly reach the bottom of the valley.

These methods can, of course, be completed and adapted according to the performed study, like for example the creation of a hazard map.

Alain Breguet et Jérôme Dubois: Landslide risk mapping in Pakistan and Nepal based on field survey and modeling techniques.

Alain Breguet et Jérôme Dubois
Co-supervisors: Prof. Michel Jaboyedoff, Karen Sudmeier-Rieux
Expert : Andrea Pedrazzini
Rapporteur : Prof. François Marillier

Himalayan region is characterized by an important density of landslides related to earthquake, monsoon rains and the important weathering of the bedrock. Several mountain communities are affected by these phenomena. A complete geotechnical hazard assessment is often difficult to complete in developing countries due to low available funding. In this study, we propose a risk mapping methodology that can be quickly applied in order to have a first overview of human risk associated with landslides in the Pakistani Kashmir and Dharan region of eastern Nepal.

This methodology proposes a quick field survey coupled with an intensive use of different types of remote sensing data. The main morphological information is collected in the field in order to have an objective analysis of the phenomena. Each landslide is mapped with its location, size and state of activity based on cracks and vegetation investigation. Participatory mapping with local communities is required to gain information and to understand which problems are caused by each landslide and to locate relict cracks and landslides. Secondly, modeling based on a digital elevation model (DEM) provides other information like a stability index (SINMAP), distance calculations and slope analysis.

The risk is assessed for buildings and roads with two different calculations. Each building receives a susceptibility level and a potential damage value based on the characterization of each landslide (frequency, volume, activity). The risk is the multiplication of these two values by the number of inhabitants per building. The roads are separated into sections below and above landslides. The calculation includes the frequency and probability of spatial occurrence of the landslide and different characteristics of the road including its length, the number of vehicles per day, vehicle speed and occupancy. Four colors are defined to show the amount of risk of each building and road section.

This methodology is quick, relatively cheap and can be applied in places with incomplete datasets. This simple methodology can be integrated into development projects by development agencies and local development authorities to consider risk due to landslide.

Guillaume Favre-Bulle: Instabilités rocheuses le long de la route nationale 7 Secteur Guido – Uspallata (km 1118 à 1133) Provincia de Mendoza – Argentina.

Guillaume Favre-Bulle
Co-supervisors: Prof. Michel Jaboyedoff, Valérie Baumann, Andrea Pedrazzini

The National Road 7, important link between Argentina and Chile, is subject to many natural hazards, threatening the vehicles travelling along. This work aims to classify the risk associated with falling blocks, to intervene as a priority on the most dangerous places to avoid investing in useless protective measures.

The study area is located west of the city of Mendoza in the Andes. The road passes several times along rocky outcrops that may be highly fractured and let blocks free. The analysis was conducted on two points: a GIS analysis of the DEM (Digital Elevation Model) and a field study. The GIS analysis has called for the creation and the search for a DEM of good quality. After several trials with satellite images and the SRTM, pairs of stereoscopic aerial photographs of 1966 have given the least bad outcome. These DEM are strongly deformed in places and do not reproduce accurately the topography.

The GIS analysis is based entirely on the study of the geometry of the DEM, the result is related to quality. Several factors are summed together : rock instabilities for the break susceptibility, slope, slope orientation, kinematic tests of the main discontinuities and SLBL. The results are used for propagations simulations of falling blocks. The results are not very good because of the poor quality of the DEM. Areas with high failure susceptibility are numerous and extensive, making it impossible to establish priority areas. The Conefall propagation can not be calibrated with field data and the results are indicative only.

On the field, discontinuities outcrops were measured and used to calculate several classification methods: RMR, SMR, and GSI. The results give rock masses of good quality, but can be very unstable following the geometry of the discontinuity and the slope of the outcrop. Then, two RHRS methods were calculated on each outcrop to classify them in order of priority for intervention. Both methods (RHRS Budetta and RHRS Colorado) take into account the characteristics of the slope, climate, geology and exposure of vehicles to danger. Both methods were compared and it is the Colorado RHRS which was recommended for future studies.

The RHRS allowed to classify outcrops for priority of intervention. Sites with the highest scores will be subject to further study to determine what protective measures will be most suitable and the money will be managed in the best possible way.

Sébastien Lévy: Analyse de l’évolution morphologique d’une partie du bassin versant de la rivière Chacoura, Québec, Canada.

Sébastien Lévy
Co-supervisors: Prof. Michel Jaboyedoff, Professeur Jacques Locat
Expert: Alexandre Loye

The large plains of Eastern Canada sensitive clays are cut by many rivers, in a way that their slopes have been and are still affected by landslides. They play a major role in the modelling of the landscape of these regions, in addition to induct a risk for human infrastructures. The role of erosion as a trigger of landslides is also important.

On the Chacoura River, north of Louiseville (Quebec), several large scars of landslides, more or less recent, are visible. A first inventory of landslides and areas of erosion was carried out in 1984 on some series of aerial photographs, covering a period from 1948 to 1979. It suggests that the majority of recorded landslides are located in areas that were previously affected by erosion. In this Master Thesis, a detailed analysis of aerial photographs, dating from 1948 to 1997 and a LiDAR-Digital Elevation Model (DEM) dating from 2007, has been done, using four different approaches: (1) a map of the phenomena on a portion of the river, which is considered as very active, was drawn by identifying various elements such as landslides, limits of the slope, position of the river, area covered by the forest and agricultural drainage structures. (2) A quantitative analysis based on the previous map has shown the temporal evolution of landslides in terms of area and volume, using statistical relationships. (3) The study focused on two flows, which happened in 1976 and 1990; it consisted of a detailed analysis of the evolution of their perimeter, area and the initial conditions of the slope. The comparison of DEMs was used to estimate volumes eroded in these flows, and in the entire zone of interest. (4) The last part focused on the general evolution of the river valley, and particularly the identification of areas where erosion could be more active in the future. This study was based on transverse and longitudinal profiles, and using the Sloping Local Base Level (SLBL) method.

The results show that the location of landslides is firmly linked to the presence of typical specific topographic features, such as (1) the shape of the meandering river, (2) the flow of agricultural drainage, or (3) the erosion at the toe of the slope. The study of landslides over a period of 59 years shows also that (4) the major landslide scars in this region are in fact the sum of several events of lesser importance. The analysis of the development of Chacoura River valley indicates that (5) it is still at an early stage of development. Besides (6) the lower part of the river seems to be relatively stable, however the upper part and the tributaries will probably face significant changes in the next 10,000 years. Finally hazard maps, based on the measured distances of flow retrogression have been created for the St-Léon-le-Grand region.