Open access
Author
Date
2019Type
- Doctoral Thesis
ETH Bibliography
yes
Altmetrics
Abstract
The mountains in the Swiss Alps are frequently subjected to extreme weather conditions, with periods of freezing temperatures, snow-melting, intense rainfall and daily temperature variations. Depending on the characteristics of the slope, geology and soil properties, several types of slope instabilities could increase the hazard to local communities. The most relevant, due to a combination of likelihood and potential mobilised volume of debris, would be during or following an intense or sustained period of rainfall that could lead to significant mass movements.
The objective of this research project, over several years, was to characterise and monitor the seasonal response of a scree slope located in the Swiss Alps (canton Valais) at an elevation of (1840-1910 m.a.s.l.). Not only are scree slopes rarely investigated in this way, they also present a number of challenges due to the deposition mechanisms in the slope and the poorly graded granular soil. The scree slope is also adjacent to an active channel that is known for having led to extensive debris flows reaching the valley below in the past. There was some concern about whether a slope failure in the scree could cause additional debris to lodge temporarily in the channel prior to release of another debris flow event.
The soil was characterised and then classified in the laboratory and stress-path dependent geotechnical strength parameters were determined. The soil response to natural meteorological events was monitored continuously during several winter and summer seasons. A schematic ground model was developed and modelled in physical and numerical simulations of surficial landslides induced by rainfall. Finally, the results are discussed in terms of the potential hazard for the community.
The project was particular by its location, which represented several challenges in terms of instrumentation and data collection techniques. The soil response was monitored through a long term field campaign, which provided information regarding soil volumetric water content (VWC) and soil temperature with depth (up to 1m) at specific locations, where trenches had been excavated and instruments had been installed. A geophysical campaign was performed simultaneously by another team member. She used Ground-based Penetrating Radar and Electrical Resistivity Tomography techniques to define bedrock depth. Additionally, the data were completed by precipitation recorded by two meteostations.
In situ and laboratory testing of the gravelly soil properties was performed, including in situ density tests and triaxial stress-path testing specific for rainfall infiltration. All these results were integrated in a ground model, which eventually served to design a prototype of the slope for investigation of the effects of groundwater and/or rainfall infiltration. Preliminary numerical modelling was then conducted using SEEP-SLOPE/W, followed by a programme of centrifuge tests in the IGT geotechnical drum centrifuge at ETH Zürich. The results showed that the soil layer is heterogeneous poorly graded gravel in terms of grain size, with some silt content. A critical state friction angle of 41° was obtained, with zero cohesion. The bedrock was located typically between 1.5-3 m depth, in a slope with inclination of 33-43°.
The soil behaviour, in terms of VWC and temperature, responded to the seasonal weather changes with a pattern: the VWC was mainly low in winter due to snow insulation and freezing temperatures, and higher and more variable in spring and summer due to the rise in temperatures with subsequent snow-melting and rainfall. The soil remained mainly in an unsaturated state because the gravel was well drained and the slope is steep. Assuming that the slope had an initial stress state ratio close to failure, the physical and numerical simulations confirmed that a shallow landslide could be triggered, depending on the combination of bedrock geometry, soil thickness, antecedent groundwater and rain intensity. Furthermore, it was found that the presence of a bedrock step outcrop that approached the ground surface is the worst scenario in terms of slope stability. A shallow landslide would be induced at lower groundwater flows than for bedrock parallel to the slope, because of the generation and rise of pore water pressure upslope of the step in the bedrock, where finally the slope failure would occur.
The study concludes from the hazard point of view that the scree slope studied in this thesis is unlikely to experience a significant slope failure, which would mobilise a considerable volume of debris that would endanger the village below; this due to the well drained conditions, shear strength of the gravel and the limited soil thickness (1-3 m). An estimate of the total debris volume overlying bedrock on this slope is 15000 m3. However a surficial landslide could be triggered in locations where the bedrock is shallow, or there is a step in the bedrock, which could force groundwater to rise following sustained rainfall and form a spring at some point. A combination of local erosion and soil saturation under a critical rainfall intensity would lead to a subsequent loss of effective stress of the gravelly soil. Show more
Permanent link
https://doi.org/10.3929/ethz-b-000377426Publication status
publishedExternal links
Search print copy at ETH Library
Contributors
Examiner: Springman, Sarah M.
Examiner: Take, Andrew
Examiner: Arenson, Lukas U.
Examiner: McArdell, Brian
Publisher
ETH ZurichSubject
Natural hazards; Landslides; Physical modelling; Numerical modelling; MonitoringOrganisational unit
03474 - Springman, Sarah M. (ehemalig) / Springman, Sarah M. (former)
More
Show all metadata
ETH Bibliography
yes
Altmetrics