Landslides in Sensitive Clays

Quick clay’s combination of sufficient undisturbed shear strength to be stable

in situ

and liquefaction upon structural failure derives from chemical factors that act during and after sediment accumulation. Flocculation of silt and clay particles in marine water produces an interlinked, random structure with a water content that approximates its high-salinity liquid limit. Co-sedimentation of cementing agents may augment its strength. Upon uplift above sea level, fresh water displacement of the high-salinity pore water decreases the liquid limit while the structure and water content remain almost constant, yielding a liquidity index in the range of about 1.5–4. The central premise of this paper is that development of solutions begins with consideration of how problems have arisen. Specifically, chemistry’s dominating role in quick clay development provides insights into approaches that can be applied to amelioration of the quick clay landslide problem.

Quantitative mineralogical analyses of sediment samples from 18 sites in Québec indicate that plagioclase is the most abundant mineral in all soils and that chlorite is generally the dominant clay mineral followed by illite and expandable clays consisting of mixed-layer clays minerals involving vermiculite. The study also illustrates how specific surface area and constitutive water content can be used to evaluate departures from average in the mineralogical composition of sensitive clays. The relationship between activity and specific surface area of sensitive clays shows that when compared to soils from different sedimentary basins that have different mineralogy, they are characterized by a much lower activity and specific surface area.

GIS Modeling of weighted quick-clay preconditions has previously resulted in clay-sensitivity predictions in SW Sweden (mainly the Västra Götaland county), but variability with depth has not been studied in connection with modeling in the area. Certain characteristics important for quick-clay distribution and related risk assessment change with stratigraphy and depth. These need to be modeled in 3D to better account for varying clay conditions. The possibilities for such work have improved with methodological advances and increased accessibility of geotechnical records. A three-step process is used to evaluate the distribution of sensitive clay within the region. First, geotechnical records are compiled and used to analyze clay sensitivity relative to their vertical distance from stratigraphically distinguishable units. Second, the results are combined with parts of an earlier quick-clay prediction model. Third, the combined results are tested against existing information and presented as 3D sensitivity. Suggestions for further refinements and applicability of findings are given based on the results.

Quick clay may be stabilised by increased magnesium concentrations in the pore water. Weathering has so far been cited as the source of magnesium, but is an unlikely process at 15 m depth in clays, where an increase was found at a test site in the south of Sweden. This study investigates how far ion exchange can explain the source of magnesium. A rough model in the program PHREEQC incorporating ion exchange, but not weathering, is used to model geochemical and transport processes since the latest ice age in a soil profile in the Göta River valley. Rain water, or rain water with added calcium or magnesium carbonate, is used as infiltrating solution. Calcium or magnesium could come from weathering in the dry crust. Advection, diffusion and ion exchange are sufficient to approximate concentrations in the pore water, if the infiltrating water contains calcium. Weathering at depths below 5 m is not included in the model, but observations of magnesium are reproduced nevertheless. The magnesium comes from the sea water at the end of the latest ice age and has been stored on the ion exchange sites in the clay. The magnesium is displaced into the pore water when calcium enters the soil, is transported downwards, and partly re-enters the exchange sites. The rough model supports the importance of ion exchange for the restabilisation of quick clay by natural magnesium at this site.

Potassium chloride was added to low saline Norwegian post glacial clays to study its effect on strength parameters. The laboratory study was carried out on undisturbed sensitive clay samples from two locations in mid-Norway. The mechanical behaviour of sensitive clays is greatly influenced by their pore water ionic content. Adding salt changes the geotechnical properties of quick clay to such an extent that it appears as a total different clay. Salt migration is a time consuming process strongly dependent on diffusion through the soil. Deriving the effective diffusion coefficient from water content of 30–50 %, the clay will be de-sensitized over a length of 50 and 56 cm respectively after 1 year. Ground improvement with salt is done by installing salt wells. This study is part of a design project for such installations. The consumption of time depends on the distance between the wells, diffusion coefficient and maintenance of high concentration in the salt well.

The Ministère des Transports du Québec is currently conducting a comprehensive inventory of large historical landslides that have occurred in sensitive clays in Québec. One hundred and eight cases have been identified during the period from 1840 to 2012 through historical documents or aerial photos. Detailed data have been collected for numerous sites with respect to pre- and post-failure topography, soil properties, the position of the failure surfaces and the groundwater flow regime. In addition, piezocone soundings have been conducted at approximately 50 other undated large landslide sites in various regions in Québec. This article summarizes the preliminary analysis of these data.

Assessment of landslides in soft sensitive clays require, in addition to geometrical aspects, a combined and complete understanding of the engineering properties of the soil materials at their original as well as disintegrated state. In light of the geotechnical characterization of 33 Norwegian landslides, this paper attempts to discuss some of the material properties which influence the extent of post-failure movements of landslides in terms of retrogression and the run-out distance. The energy involved in remolding sensitive clays has been studied for the Norwegian sensitive clays to signify its importance with regards to the estimation of extent of landslides.

The dimensions of landslides are subject to geometric, material and external controls. Geometric controls include: slopes and orientation of the ground surface and bedding planes, valley width and geometry, depth of failure, plus the presence and incision depth of bounding streams. Material controls include undrained and remoulded shear strength, sensitivity, stability number, remoulding index, rapidity, and the mobility of the depleted mass. Dense networks of trees can influence mobility. The position of layers prone to landsliding in the slope is important.

The ages of 39 large retrogressive landslides and three areas of disturbed terrain in the Ottawa Valley were compiled using 122 published and unpublished radiocarbon ages. The chronological dataset includes 15 confined-valley landslides, 23 scarp-side landslides, and four special case features (one massive landslide, three disturbed terrain areas). The ages of the features range from ‘modern’ to 8,000

14

C cal BP. Distinct clusters of 10 and 11 coincidentally-aged landslides at ~1,000 and ~5,150

14

C cal BP are the groups of landslides previously interpreted to have been triggered by paleoearthquakes. Scarp-side landslides with scars between 0.1 and 10 km

2

are the dominant failure morphology forming the two age clusters and constitute an important component of the interpreted paleoseismic evidence. Five of the confined-valley landslides are part of the ~1,000

14

C cal BP cluster, but the other ten failures are of widely varying ages. One of the special case features, a massive landslide originating from a source area of ~20 km

2

, falls within the ~1,000

14

C cal BP cluster. Radiocarbon ages representing the age of the Treadwell and Wendover disturbed terrains, suggest that these areas are contemporary with the Lefaivre disturbed area at ~7,900

14

C cal BP, but this is not an unequivocal interpretation. Notwithstanding sampling bias within the dataset, it is inferred from the high proportion of dated landslides falling within the two interpreted paleoearthquake clusters that there is a ‘strong’ paleoseismic signature within the temporal pattern of landsliding within the Ottawa Valley. Based on the Ottawa Valley dataset, scarp-side landslides with preserved debris fields and scars greater than 0.1 km

2

, and landslides in general with scars greater than 1 km

2

, are more promising targets than confined-valley landslides for paleoseismic studies.

Early in the morning of April 3rd, 1980, approximately 1.15 × 10

6

m

3

of soils from the coastal bluffs spread out over the tidal flat along the North Shore of the Jacques-Cartier Strait in the Gulf of St. Lawrence. This landslide had a width of 410 m and had retrogressed back up to 110 m from the slope crest, cutting through Highway 138, 9 km west of the town of Havre-St-Pierre. In this area, the coastal bluffs are approximately 23 m high and consist, from the bottom to the top, of a thick marine sensitive clay unit from the Goldthwait Sea, overlain by 3 m of sand and 3 m of peat on top. The debris morphology has a “thumbprint-like pattern” characterized by unbroken blocks of intact clay forming elongated ridges, surrounded by a mixture of remoulded clay, sand and peat, aligned parallel to the bluff and arched in the direction of the flow. The absence of obstacles and the non-channeled character of the flow path of the Havre-St-Pierre landslide provide a good opportunity to characterize and analyze the post-failure stage of this large landslide in sensitive clays.

In this paper, we present the geotechnical properties and microstructure of a clay from Mosjøen, Northern Norway, where a landslide occurred in June 2011. The evolution of the clay properties at the depth of failure (i.e. shear zone) are studied based on an integrated set of samples, laboratory tests, CPTU data and piezometer readings. Results from SEM analyses show a general lowering of the macro porosity of the clay due to shearing during the landslide. As a consequence, the lower permeability within this zone slowed down the consolidation process when compared to the surrounding intact material. Despite of this, as pore pressure decreased from a situation with nearly zero effective stress to a situation close to the hydrostatic pore pressure, the shear resistance of the clay slowly increased; from its remoulded strength and towards values slightly higher than typically found for normally consolidated clays. It was also found that the peak strength envelope of the destructured clay was above the envelope of the intact material. The reasons could be attributed to the microstructure of the remoulded material and to its silty nature.

Marine clay deposits in coastal, post-submarine areas of Scandinavia and North America may be subjected to quick clay landslides and hence significant efforts are being taken to map their occurrence and extent. Recently, considerable efforts by a number of researchers have been made to investigate areas of sensitive clay using a range of geophysical techniques. Although the majority of this work has focussed on measurements of electrical resistivity, other electromagnetic and seismic geophysical techniques have also received attention in the literature. The purpose of this paper is to review recent research concerning the effectiveness of a number of geophysical techniques for investigating sensitive clays. In addition to discussing a number of case studies, this review will also consider recent work showing the correlation of geophysical measurements, and in particular electrical resistivity, with a range of relevant engineering properties.

During the last 10 years, several Norwegian projects have explored the possibilities of using 2D resistivity measurements for quick-clay mapping in combination with traditional geotechnical methods. Experience has shown that an effective first-order interpretation of 2D resistivity profiles can be performed by using the following classification: unleached clay deposits (1–10 Ωm); leached clay deposits, possibly quick (10–100 Ωm), and dry crust clay deposits and coarse sediments (>100 Ωm). However, resistivity values are influenced by local conditions and there is an overlap between the classes. The 2D resistivity method can prioritize areas for further investigation using other geophysical methods or drilling to refine the interpretation of the subsurface. Through several case-studies, the 2D resistivity method has proven useful for detecting potential layers of quick clay, for outlining the extent of these layers and their positions in slopes (also near the shoreline), and for engineering applications such as construction planning. To this end, the method has shown to be applicable for adjusting the extent of hazard zones to improve stability evaluations. Another important application is landslide investigations, to identify barriers that may deter further landslide propagation.

Quick clay is highly sensitive, marine clay with an unstable mineral structure due to post-glacial heaving and subsequent leaching of saline pore fluids by surface- and groundwater. Quick-clay layers pose a serious geo-hazard in Scandinavia and North America and need to be delineated in detail. Geophysical methods, especially resistivity methods, have been tested for quick-clay mapping at several sites across Norway. By scrutinizing results from Electric Resistivity Tomography (ERT) and integrating them with geotechnical borehole data including Resistivity Cone Penetrometer Testing (RCPT), we confirm the value of an integrated study for quick-clay hazard zonation. ERT is an ideal tool to interpolate limited borehole results and thus to provide a more cost efficient and detailed result than with borehole data alone. Our resistivity data from ERT, RCPT and lab measurements are generally consistent and appear isotropic. Geochemical analysis confirms that changes in resistivity are directly related to changes in clay salt content and secondarily related to sensitivity. It’s a challenge to resolve small contrast in resistivity (to distinguish unleached from leached clay), in close vicinity to drastic changes in earth resistivity, for example the transition from clay to bedrock. To cope with this we improve results by means of constrained ERT inversion approaches based on drilldata and/or seismic refraction bedrock data. Though ERT is no silver bullet solution to detailed quick-clay mapping, it can provide a significant contribution to improve the risk assessment at comparably lower costs than extensive drilling campaigns. Remaining methodological ambiguities need to be handled by integration with further data.

The paper presents results from site investigations for a major public highway project close to Trondheim in Mid Norway. The planned road is located in an area with thick deposits of sensitive and quick clays, and is expected to represent both local and global stability issues. The site investigations included 2D resistivity measurements and conventional geotechnical borings, such as rotary pressure soundings, total soundings, cone penetration tests (CPTU), undisturbed piston sampling and pore-pressure measurements. Extensive laboratory investigations were also carried out, including salinity measurements. The 2D resistivity measurements were carried out to detect the distribution of leached clay, and the results were used for planning of the subsequent geotechnical investigations. This is one of relatively few Norwegian studies using 2D resistivity measurements in evaluation of ground conditions in an industry project. There was very good agreement between the data sets with respect to presence of sensitive or quick clays. The investigation hence clearly points out that the combination of resistivity measurements and geotechnical borings has the potential of becoming a powerful site investigation strategy, particularly in mapping of large areas or long-stretched road corridors.

Within clay prone areas it is important to know if and to what extent quick clay is present. In Sweden undisturbed sampling and laboratory investigations are used to determine clay sensitivity and identify quick clay. Recent research for mapping of quick clay areas with the resistivity method is based on the fact that low pore-water salinity constitutes a criterion for quick clay formation. Electrical resistivity tomography (ERT) is an established geophysical method that provides an overview of the resistivity in large volumes, but resolution decreases rapidly with depth. CPT-R probes (Cone Penetration Test with Resistivity module) on the other hand can register the resistivity with very high resolution continuously along the probed depth. In a known quick clay area along the Göta river in Gothenburg, Sweden, both ERT and CPT-R were used to evaluate possible synergy effects when the results were combined. Field investigations comprised 11 CPT-R soundings and five parallel ERT profiles, each 400 m long. The results gave good agreement between the methods down to depths of 10–15 m. Below that the CPT-R probe generally registered lower values than the ERT inverted resistivity models. The results from ERT comply with the value of 6 Ωm as a lower threshold value for possible quick clay formation in Sweden. From CPT-R measurements this value was found to be approximately 3 Ωm. The discrepancy appears to be related to sulphide content and may have measurement technical explanations or be due to the differences in chemical composition. The results show that a combination of ERT and CPT-R provides an efficient approach for mapping of possible quick clay volumes. It can be an integral part of an optimized detailed geotechnical investigation program, and form a decision basis for where to do undisturbed sampling for laboratory analyses. It also opens possibilities for correlation between resistivity and mechanical parameters via the CPT-R results.

Quick-clay landslides are a known hazard in formerly glaciated coastal areas. Some of Norway’s most densely populated areas are located in potential quick-clay zones and, hence, large efforts are devoted to map the distribution of quick clays. Here, we focus on one particular Norwegian site (Hvittingfoss, 100 km south-west of Oslo), which was remediated against potential sliding in 2008. A set of geophysical methods including Electrical Resistivity Tomography, P-wave seismic refraction tomography, S-wave seismic reflection profiling, and Ground Penetrating Radar, were jointly analysed and complemented with laboratory data and

in-situ

geotechnical measurements (i.e., CPTU, SCPTU and RCPTU) in order to establish a suitable, integrated and multi-disciplinary approach to map the extent of the quick-clay zone. Through careful integration and interpretation of the different data, the main deposits were identified. Both the clay deposit and the overlying sand layer were precisely imaged and their lateral variations were determined. The underlying moraine deposit and the bedrock were also identified, thereby yielding an idea of the preferential leaching paths. Considering the inherent complexity of quick-clay mapping, the collected data illustrate the benefit of an integrated approach, and emphasise the need for high resolution, proper imaging, calibration and ultimately joint inversion of the different data.

The joint acquisition and processing of vertically polarized shear (

SV

) wave seismic reflections and surface waves during a seismic survey were carried out in Buckingham (Québec), near Ottawa, Canada, to characterize a thick (20–40 m) sensitive clay deposit. At the study site, the outcropping clay unit overlays a 20–50 m thick layer of sand and gravel and the bedrock depth reaches more than 90 m along the survey line. The seismic reflection survey using common-mid-point (CMP) inversion of

SV

-wave reflections allowed the localization of the clay-sand and sand-bedrock interfaces as well as the measurement of

SV

-wave velocities down to the bedrock contact. Velocity variations at depths less than 10 m could not be assessed due to the early reflections hidden by seismic arrivals such as surface waves. However, multi-channel analysis of surface waves (MASW) provided the variations in

S

-wave velocity from the surface down to a depth of 12 m at each CMP location. The joint acquisition and processing of

SV

reflections and Rayleigh waves provided a more complete and accurate 2D

SV

velocity model than both methods taken separately. To test the accuracy of the proposed approach, a multi-offset seismic piezocone penetration test (SCPTu) was performed along the survey line from the surface down to a depth of 25 m. The vertical variations in seismic velocities in sensitive clay as inferred from the

SV

seismic reflection survey and MASW are comparable to the SCPTu

S

-wave profile.

Geophysical and geotechnical data are presented from different sites in eastern Ontario where variable geotechnical properties of Champlain Sea sediments (‘Leda Clays’) are found. Sites range from thick “undisturbed” silts and clays, to “disturbed” geologically similar soils (earthquake triggered landslides and other deformed materials). High-resolution seismic profiles provide stratigraphic context for some of the boreholes drilled in the study area. Downhole geophysical logs from 14 boreholes are compared to core sample measurements of porosity, sensitivity, and porewater conductivity to develop useful empirical relationships. According to these relationships, silt and clay sediments can be sensitive or quick when formation conductivity drops below 100 mS/m. Conversely, silts and clays with elevated conductivities (>250 mS/m) are rarely sensitive. Salinity values calculated from porewater conductivity indicate sensitive or quick behaviour may be expected in leached soils when salinity drops below 2 g/l.

For understanding the processes acting during the triggering and evolution of a landslide in sensitive clay, it is of importance to account for the material response in the failed state. This includes the rate of strain softening and the level of residual resistance. Through a review of new experimental data it is illustrated here that these properties depend on the displacement rate in the case of shear band formation. Higher rate increases the brittleness both in terms of a steeper softening response and a lower residual level. These observations are attributed to processes of local drainage and dissipation of pore pressure gradients in the vicinity of the shear band. Viscous effects, which yield higher resistance for higher rates of deformation, are also apparent. These aspects of material failure in sensitive clay are utilized in the context of progressive slope failure in order to explain observations from case records and to quantify the impact of the rate dependent softening response on the resistance of a slope. The strong brittleness of sensitive clay at high displacement rates under shear band formation motivates a recommendation of using a design approach close to a first yield criterion if the construction situation might include such high rates.

Through detailed case studies from the literature it is suggested that a sensitive clay spread is formed by propagation of a failure surface in an intact slope and dislocation of the soil mass in horsts and grabens. It is proposed that the initiation and propagation of the failure surface can be explained by progressive failure mechanism. According to this failure mechanism, failure is initiated near the toe of the slope and the strain-softening stress-strain behaviour of sensitive clays is used to redistribute shear stress along the quasi-horizontal shear zone. The failure propagates inside the deposit reducing the horizontal stress. Active strength of the soil may be mobilised, explaining the dislocation of the soil mass above the shear zone in horsts and grabens. A numerical procedure is used to back calculate the 1994 spread at Sainte-Monique, Québec, Canada, involving slightly over-consolidated sensitive clay. The initiation and extent of the failure surface observed on site are explained by a soil having large brittleness during shear and large-deformation shear strength close to the remoulded shear strength of the soil.

Many geomaterials exhibit a reduction in strength after attaining their peak strength states. This behavior, also known as strain softening, has commonly been attributed to reductions in the cohesion and/or friction angle of the materials. This holds true especially when these strength parameters are evaluated at a very large strain levels. However, for the strain levels achieved by standard undrained triaxial tests on soft sensitive clays show that the friction angle and cohesion of these materials remain almost unchanged, even when they display a significant reduction in post-peak shear strength under undrained condition. In such conditions, the increase in shear-induced pore pressure is observed to be responsible for the observed undrained softening in soft sensitive clays. This paper elaborates on this aspect based on the behavior of soft sensitive Norwegian clays.

The effect of strain softening in geotechnical design of fills in areas with soft sensitive clays is studied by a large number of finite element analyses. The reduction in undrained shear strength with increasing shear strain after the peak value will reduce the maximum fill height before failure compared with a perfectly plastic material. The finite element program Plaxis together with the material model NGI-ADPSoft are used in this study. A non-local strain formulation is used in NGI-ADPSoft to overcome the crucial problem of mesh dependent results typical for this type of problems. The effect of brittleness is then fully controlled by input parameters. The material properties are taken from NGI’s database of undrained shear test results on high quality block samples. The effect of strain-softening is quantified by establishing a scaling factor F

softening

that gives the ratio between the calculated capacity without and with the effect of softening. The purpose is then that the peak undrained shear strength of sensitive clays simply can be divided by this factor before used in conventional limit equilibrium analyses with a strain independent (perfectly plastic) assumption to indirectly account for the effect of brittleness.

This article discusses undrained effective stress stability analyses on normally consolidated soft clay. A key issue in such analyses is that the yield induced pore pressure is properly accounted for. Yielding of natural clays is affected amongst other factors on the shape and size of an anisotropic yield surface and the deformation properties of the clay, including creep. These then have a large effect also on the undrained capacity of the soil. The implementation of yield induced pore pressure into stability analyses conducted by both limit equilibrium method (LEM) and finite element method (FEM) is presented. Calculation examples are given for a full scale embankment failure experiment using various methods to include the yield induced pore pressure. It is shown that it is possible to account for such effects even in simple LEM analysis to obtain a realistic failure load, while in advanced FEM analysis it is possible to also account for time effects.

The majority of the population of Québec (Canada) is settled in the St. Lawrence River Lowlands, where soils consist mainly of clays. In some areas, these clays are very sensitive, and are prone to the development of large retrogressive landslides. To identify areas that are potentially exposed to this hazard, the Québec government produces maps of areas prone to landslides in clayey soils. These maps allow the government authorities to take this risk into account in their operations. Furthermore, municipalities must include this mapping in their regulations for land use planning. The risk analysis for landslides in clayey soils is based on these maps. Different approaches are currently being developed in order to manage risks according to the types of landslides and the elements at risk (population, roads, etc.).

Parts of the Norwegian roads and railway network are located in areas dominated by sensitive clays. Several new projects are facing challenges in areas with deposits of sensitive clay. The design codes and guidelines of the Norwegian Public Roads Administration (NPRA) and the Norwegian National Railways Administration (NNRA) have been developed over many years with respect to ground investigations and safety requirements. These regulations are specific for road and railway construction. In addition there are general design codes like the Norwegian Standards, and the regulations established by the Norwegian Water Resources and Energy Directorate (NVE) concerning the safety of larger areas adjacent to e.g. roads and railways. This paper points out the manner in which the regulations impose an increased attention to safety when sensitive clay is encountered. The need for further development of the regulations, based on recent studies and research, is also described.

The approach for the assessment of the risk associated with quick clay slides in Norway is a qualitative/semi-quantitative procedure developed as part of work for The Norwegian Water Resources and Energy Directorate. Slide areas are classified according to “engineering scores” based on an evaluation of the topography, geology and local conditions (to qualify hazard) and an evaluation of the elements at risk, persons, properties and infrastructure exposed (to qualify consequence). The risk score to classify the mapped areas into risk zones is obtained from the relationship

R

S

= H

WS

× C

WS

, where

R

S

is the risk score,

H

WS

is the weighted hazard score and

C

WS

is the weighted consequence score. The risk matrix is divided in five risk classes. Guidelines for the implementation of the risk matrix are administered by NVE. In practice, the approach is used to make decisions on required mitigation measures to reduce the risk. The approach is simple and makes room for engineering experience and judgment. For detailed regional planning, slope stability calculations need to be made. Methods for quick clay slide stability calculations taking into account the brittle behaviour of the material are under development. This chapter provides an illustration of this development work.

For decades, mapping of Quaternary geology in Norway has been an important basis for hazard and risk assessment for landslides in highly sensitive clay (quick clay). One aspect of particular significance is information on the occurrence of fine-grained, marine deposits including clays. The marine limit (ML) defines a natural upper limit to these deposits, so issues concerning marine clays can be disregarded above ML. Below ML, a filtering of the Quaternary map information is needed to identify areas where clays are potentially present. Fine-grained, marine deposits are more frequently encountered below some deposit types than others, and the possibility of encountering marine clays within an area is defined as clay-deposit susceptibility. Stratigraphic information is needed for verification. Further landslide susceptibility assessment requires additional information on topography and ground conditions. Other geological information is also important such as on the distribution of landslide debris and landslide scars. Here, it is beneficial to use high-resolution, digital elevation models for detailed analysis of the terrain also below sea level in near-shore areas. Maps with filtered geological information can work as a supplement to maps produced during the ongoing quick-clay mapping program and as a help prioritize areas requiring further investigation.

In order to limit the vulnerability of society to effects of the climate change, the Swedish Government in 2008 commissioned the Swedish Geotechnical Institute to map the risk of landslides along the Göta River. The Göta River valley is one of the areas most prone to landslides in Sweden. Since it is an area with highly sensitive clays, mapping of quick clay areas and estimation of possible landslide extent were important parts of the investigation. In the Göta River Commission, a correlation between the sensitivity and the total rod friction evaluated from CPTs and static pressure sounding tests was used as a complement to sampling for mapping of quick clay areas. If a landslide is initiated in an area with quick clay the probability is high that it will propagate and affect areas nearby. Within the Göta River Commission, estimates of how far a landslide would extend backwards was made taking into account the slope geometry and the sensitivity of the clay, partly based on analyses of earlier occurred landslides. A methodology was developed for rational management of quick clay areas in the stability investigations for the Göta River valley.

The use of partial safety factors has, with the adaptation of the Eurocodes, become the dominant safety concept, also for slope stability in most European countries. The article discusses the partial safety factor approach based on considerations of uncertainties involved and consequences of failure. Partial safety factors are calculated based on reliability theory, following first the present ideas of the Eurocodes, and second an alternative approach aimed to improve the shortcomings in the code. In the alternative approach, all uncertainties are placed on the material safety factor, which are then calculated for different target reliability index values. Two simple examples are presented, where the differences of the two approaches are outlined.