Landslides in Sensitive Clays

Landslides in Eastern Canadian sensitive clay deposits are generally located along river or stream valleys. The stratigraphy of these deposits can be simplified as an impervious clay deposit between two pervious boundaries, a weathered fissured crust at the top and a coarse till layer at the bottom. The paper examines the evolution of the ground water regime as the valleys are formed by erosion and discusses the impact of the valleys formation on the requirements for slope stability analysis and determination of shear strength. A large data bank of laboratory test results are then presented and treated to arrive at a determination of the shear strength parameters based only on the preconsolidation pressure.

The property of silty-clay to clayey-silt quick clays, whereby apparently solid soil transforms to the liquid state when subjected to sufficient stress, derives from chemical factors: mineralogy (low activity); depositional environment (marine-to-brackish conditions causing flocculation and high water content); and post-depositional chemical changes (development of cementation and displacement of marine-to-brackish water by infiltrating rainwater). The stability of slopes developed by river incision is affected (negatively) by physical factors (drainage and fluctuating water tables) and chemical weathering reactions that have led to weak, fissured, blocky, nodular structures. Immediate causes of quick clay landslides are commonly the physical factors of: river erosion; high water contents in the fissured slopes; and human actions. Regardless, the characteristics of the resulting landslides are primarily determined by the chemical factors.

Installation of salt wells filled with potassium chloride may be used as a ground improvement method in quick-clay hazard areas. The migration of potassium chloride is self-driven by molecular diffusion. The effectiveness of improving the geotechnical properties and the time to do so, depend on hydrogeological conditions at the site, original pore water composition and concentration, adsorbed ions, mineral type, and cation exchange capacity. Increased salt content in the pore water decreases the repulsive forces acting between the mineral surfaces. By decreasing the repulsive forces, the liquid limit and remolded shear strength increase, indicating the improvement of the post-failure properties. The clay particles no longer repel one another, ultimately preventing development of retrogressive landslides.

When dealing with slope stability considerations in deposits where sensitive and quick clays might be encountered it is vital to map the extent of these clays. For the geotechnical engineer, the cone penetration test with pore pressure measurement (CPTU) is a powerful tool in this respect. With its combined measurement of tip resistance, pore pressure and sleeve friction, the CPTU holds a great potential for identification of quick and sensitive clays. Such interpretations can be done based on measured data directly or by combining parameters in dimensionless numbers. Amongst the more popular dimensionless numbers are the pore pressure ratio (B _q ), the cone resistance number (N _m ) and the friction ratio (R _f ). Diagrams exist which allow classification of soils based on the combination of such numbers. Robertson (Can Geotech J 27:151–158, 1990) is one widely used example. However, In Norway, it is found that existing diagrams to a large extent fail to identify sensitive and quick clays. Based on a database of 10 Norwegian sites a new set of classification diagrams are presented with focus on identifying quick and sensitive clays. The diagrams are based on a pore pressure ratio where the tip pore pressure is used (u ₁ ) rather than the u ₂ -position as this is found to better capture the actual collapsible response of sensitive clays. The cone resistance number is modified to also include an effect of overconsolidation (OCR) instead of only accounting for vertical effective overburden. Also, the friction ratio is normalized with pore pressure (u ₁ ) rather than the cone resistance. Electrical resistivity values from R-CPTU-soundings are also included in the considerations. The outcome is a set of revised classification diagrams that provides more accurate identification of Norwegian sensitive and quick clays compared to existing classification diagrams.

Shear wave velocity (V_s) is an important parameter in various geotechnical investigations especially in dynamic problems but is also useful in general characterisation studies and for problems such as landslide hazard assessment. This paper presents V_s data for sites in Norway and Sweden and shows that V_s can be measured reliably and repeatability using a variety of intrusive and non-intrusive techniques. Norwegian and Swedish V_s profiles show similar trends with depth. However the values for Norwegian soils are much higher. When correlated against the index properties water content and density the reason for the differences can be seen and the correlations apply across the soils from the two areas. Good correlations are found between V_s and s_u and V_s and p_c′ and it is also shown that one can predict V_s from CPTU data. However it is not possible to use correlations developed for Norwegian soils for Swedish clays and the results of this work show that locally derived correlations are needed. This work also shows that it is not possible to predict whether a clay is quick or not from V_s measurements alone.

The results of a geophysical and geotechnical investigation in a sensitive clay deposit affected by numerous landslide scars in Vases Creek Valley near Brownsburg, Quebec, Canada are presented herein. The main objective of this investigation was to assess the suitability of electrical resistivity measurements in marine clay deposits for mapping out areas prone to flowslides. In addition to a 1.6 km-long electrical resistivity tomography (ERT) carried out perpendicular to the axis of the Vases Creek Valley, six piezocone penetration tests and five boreholes with sampling were also performed along the geophysical survey line. Moreover, standard geotechnical parameters and pore water salinity, as well as electrical resistivity of undisturbed clay samples were measured in the laboratory. According to the correlations found between the remoulded shear strength, the pore water salinity and the electrical resistivity, clay samples with salinity below 6.2 g/l are characterized by remoulded shear strength below 1 kPa and electrical resistivity above 2.8 and 10 Ωm measured respectively in the field and in the laboratory. In such conditions, sensitive clay deposits can be prone to flowslides if all other criteria are also met. Based on this resistivity limit value, only one small area of non-sensitive clay was identified in the interpretative stratigraphic cross-section assessed from the field investigation. Otherwise, the deposit is entirely composed of sensitive clay. The ERT is a promising geophysical tool for the delineation of areas prone to large landslides in eastern Canada.

This laboratory study involves leaching clay from Onsøy, Norway. Deaired deionised water reduced the pore water salinity, potentially forming a quick clay, in a triaxial cell, modified to allow shear wave velocity and resistivity measurements to be made. This project aims to assess how changes in the geotechnical properties of the clay influence its geophysical properties. The testing procedure has been able to create a quick clay with a remoulded shear strength of 0.2 kPa, and a final salt content of 2.0 g/l. This corresponded to an increase in the resistivity of the clay from initially 0.9 Ωm to a final resistivity of 14.0 Ωm.

Energy involved in disintegrating of sensitive clays from an intact to a fully remoulded state is one of the key aspects in assessing the post failure movements of sensitive clay landslides. This energy is referred to as remoulding energy. In this paper, the energy approach is conceptualised using an analytical approach. A comprehensive review of the empirical, laboratory and field techniques to estimate remoulding energy are presented and discussed in detail.

In Finland, undrained shear strength is commonly measured using the field vane shear test (FV). Currently, the most commonly used field vane testers are the Nilcon vane and the electrical vane with shear rotation and measuring systems located above the ground level. Vane testing is normally carried out using vanes equipped with slip coupling, while the use of casing for protecting the vane is not very common. Recent studies from Finland have shown that the undrained shear strength of clays can be significantly underestimated when casing is not used. Experimental observations suggest that the slip coupling might not always be sufficient to remove all of the rod friction effects that occur during testing. Tampere University of Technology has recently purchased an innovative field vane apparatus with a vane tester unit, where torque and rotations are measured right above the vane. In this way, the effect of rod friction is minimized and the measured stress-rotation behavior is less biased. In this study, issues related to practical applications, testing devices and interpretation methods are discussed. Then, a critical comparison between test results in soft clays from both the traditional and new field vane testers is performed.

During block sampling the in situ total stresses reduces to zero. This ultimately allows the soil sample to swell, leading to a weaker soil structure. In this paper, an attempt has been made to investigate this mechanism experimentally. In doing so, a new laboratory test procedure has been developed where the formation of a soil is simulated with a built in piezometer to study the stress changes in soil samples during and after sampling. The results show that the tested sample tends to lose a significant part of its residual effective stresses instantaneously, allowing the sample to swell.

A large part of the uncertainty in geotechnical design originates from the soil properties, which are assessed through the site investigation. The use of high-quality soil block sampling may reduce these uncertainties, but such methods are only available for comparably shallow soil layers. Deeper soil layers are of interest for analysis of large progressive landslides, where the slide surface can reach a significant depth (>20 m), and where the assessment of the stability requires information about the strength, and in some cases the compressibility, of the soil. A case study of sample disturbance in such deep layers of clay is therefore discussed to highlight how sample disturbance can be detected in different types of measurements from a site investigation.

The stress-strain response of sensitive clays tested in a laboratory setting can be significantly affected by disturbance effects caused by sampling, transport, storage and specimen preparation. Soil models for finite element analyses are commonly calibrated using the results from laboratory tests and, consequently, calibrated model parameters are likely to be affected by sample disturbance. For sensitive clays subjected to constant volume shearing, the stress-strain behavior is dependent on the direction of loading and, due to build-up of shear induced pore pressure, effective stresses will reduce with increasing strain in the post-peak regime. According to previous studies, peak strengths, strains at failure and post-peak behavior of sensitive clays are all significantly influenced by sample quality. Therefore, the relative quality of model predictions generated using a sensitive clay finite element model can also be expected to be notably affected by sample disturbance. In this study, the impact of sample disturbance on the determination of model input parameters for advanced finite element modelling of sensitive clays is addressed and critically discussed. Two advanced soil models are used for this purpose: the total stress based NGI-ADPSoft model, which is able to predict the anisotropic strain-softening behavior of saturated sensitive clays, and the effective stress based S-CLAY1S model, which is characterized by an anisotropic yield surface and is able to simulate soil destructuration. The practical implications of a thoughtful selection of the input parameters are evaluated through FE stability analyses of a sensitive clay slope.

The flow behaviour of fully remoulded sensitive clay from Esp, near Trondheim, Norway, is investigated using a wide-gap concentric-cylinder viscometer. The salinity S and liquidity index I _L are varied over a wide range. The Esp clay shows shear thinning flow behaviour, similar to Canadian sensitive clays. The Herschel–Bulkley (H-B) model allows a better fit to the measured flow curves than the Bingham model for the Norwegian as well as Canadian clays. All model parameters depend strongly on the liquidity index and the salinity. The H-B exponent n was found to vary from 0.36 at high S and low I _L to 0.97 at low S and high I _L . The onset of turbulence in the rheometer is detected at Reynolds numbers of 50–100.

This paper describes a detailed geotechnical site investigation carried out to study the static and dynamic properties of a post-glacial Goldthwait Sea sensitive clay deposit. The focus is on the evaluation of the small-strain shear modulus of the sensitive clay. The site is located on the North Shore of the St-Lawrence Gulf (Province of Quebec) and it is selected as it is the location of a major landslide triggered by rock blasting operations in 2009. The investigations involved both in-situ and laboratory tests (resonant column and bender element testing). The laboratory tests were carried out on high quality large diameter clay samples from three different depths. The objective of this paper is to compare the laboratory small-strain shear modulus obtained with resonant column and bender elements with the small-strain shear modulus obtained in the field using the seismic piezocone (SCPTu) and ambient vibrations.

The mechanism of initiation or triggering of landslides in soft sensitive clays is governed by strain softening and localization of deformation into thin shear bands. The triggering load and the extent of the initial part of the landslide depend upon the formation and propagation of these shear bands and in particular upon their thickness. The role of local pore water drainage and rate dependency in the formation and propagation of these shear bands is not fully understood. This paper reviews some of the research carried out during the last decades within this field, discuss the results and points to some recommendations for further research in order to reach a conclusion.

Vibratory rollers transmit vibrations that may lead to pore pressure build up and soil failure in vibration susceptible soils. In a recent incident at Statland in mid Norway, a vibratory roller had compacted a shoreline embankment fill shortly before a submarine landslide was initiated. The landslide triggered a tsunami causing economic damages for several millions NOK. This incident shows the need for guidelines for controlling the effects of construction activity induced vibrations on slope stability. In a follow up study, the effect of vibratory roller compaction on soil degradation has been evaluated with a finite element model. The effect of vibrations varies with parameters such as soil stiffness, bedrock depth and geometry, and presence of thin soft layers. To develop a vibration measurement procedure, we suggest, based on a threshold shear strain of 0.025%, a 15 m wide and 5 m deep influence zone outside of which the soil strength is not reduced. A tentative vibration limit is set to 10 mm/s near slopes with vibration susceptible materials. We also provide some recommendations on how to account for vibratory compaction in slope stability analysis.

Safety assessment of natural slopes in sensitive clays is subjected to uncertainty due to the natural variation of soil properties, measurement and modelling errors. In order to ensure acceptable safety levels, geotechnical design codes (e.g., Eurocode 7) commonly provide frameworks for a systematic treatment of uncertainties in the safety assessment of a slope. The treatment of uncertainties in the design codes is primarily focused on the parameters directly involved in the analysis of the mechanical stability of a slope (e.g., soil strength parameters). Additional valuable contributions to the safety assessments of slopes can be also provided by information that relates indirectly to the mechanical stability of a slope (e.g., past slope performance). However, there is often a lack of systematic integration of indirect information in the existing design codes. This paper examines the integration of indirect information based on the observed past slope performance in the safety assessment of a slope. The integration is facilitated through the Bayesian framework because it provides a basis to update uncertainties in the slope stability and safety assessment, such that they are consistent to the observed slope performance. The paper examines the effects of slope survival and failure events on uncertainties in the slope stability analysis.

The North Spur is a piece of land acting like a natural dam whose stability is crucial for the Muskrat Falls hydroelectric project. Stabilization measures have been designed to improve its stability taking into account the presence of sensitive clay. This paper discusses the potential landsliding at this site and presents the methodology used to study the progressive failure mechanisms which could impact the stability of the North Spur.

Correction factors to be used in conventional undrained stability calculations in order to account for post peak strain softening behaviour of sensitive clays, has been recommended based on an extensive sensitivity study with advanced finite element simulations. It is found that a correction of the material factor is preferred compared to a reduction of the shear strength. The input parameters to the sensitivity study that had the highest correlation with the required correction factor were the shear strength increase with depth and the brittleness, which is the rate of shear strength reduction with strain. In a large block sample database of sensitive Norwegian clays, there was no clear correlation between the brittleness and the sensitivity. Hence, classification of the clays based on the sensitivity is not recommended for evaluating the effect of strain softening on the capacity.

It is well known that in undrained stability calculations, total stress and effective stress analyses do not give the same calculated factor of safety when FOS >1. This is due to the fact that shear strength is defined differently in these two approaches: In total stress analyses, the mobilised shear stress is compared to undrained shear strength, i.e. strength at failure. In undrained effective stress analyses, the shear strength is defined as corresponding to the mobilised effective stress state. This causes an overestimation of FOS in undrained φ′-c′ analyses. Modelling of excess pore pressure Δu has traditionally been source of most uncertainty in undrained effective stress analyses. Having the correct shear strength along the slip surface can be considered the most crucial detail in all stability analyses. It can be argued that in the context of Limit Equilibrium analyses where deformations are not considered, priority should be given to calculating the shear strength correctly, instead of attempting to obtain a “correct” mobilised Δu value. This paper gives a general introduction to the new HSU (Hybrid s _u ) method. For the purposes of LEM analyses, Δu is calculated so that the resulting Mohr-Coulomb shear strength corresponds to the assumed failure state. This approach solves the inherent overestimation of FOS in undrained φ′-c′ analyses. To predict the effective stress at failure, a constitutive effective stress soil model is employed. Also presented is a concept of deriving undrained shear strength s _u in LEM, based on an effective stress soil model. This makes it possible to conduct the LEM stability analysis in terms of total stresses, while deriving soil strength from effective strength parameters. The different approaches of calculating Δu and s _u with the HSU method are compared using a theoretical stability calculation example. The relative merits of the different approaches are discussed.

This paper presents a recommended practice for the use of strength anisotropy in stability calculations. The recommendations are based on laboratory data from high quality block samples collected by NGI from more than 20 sites. The strength anisotropy was correlated against natural water content, OCR, sensitivity, plasticity index and clay contents. Despite some scatter in data, the paper presents correlations to estimate strength anisotropy for Norwegian clays. A benchmark stability calculation has been done to illustrate the overall impact of various anisotropy factors on the factor of safety.

During construction of E45 – Norge/Vänerbanan, analyses of an area located in Agnesberg, just north of Gothenburg, showed an evaluated factor of safety significantly less than prescribed in the codes for both the new main road E45 and the new railway Norge/Vänerbanan. The road and railway are situated very close to each other, just next to the Göta River that complicates the evaluation of the safety factor. The empiricism commonly used was questioned due to an unfamiliar change in undrained shear strength in the profile. Consequently, an extensive field and laboratory campaign was conducted to establish a more accurate and reliable strength profile. The extensive laboratory programme contained a number of direct simple shear tests, as well as triaxial tests in both compression and extension to be able to determine the strength anisotropy in the area. The laboratory study showed that the soft clay in the area was more anisotropic than anticipated, which was more favourable in the studied case regarding the stability. This lead to a new anisotropic function to be used in evaluation of the safety factor. This combined with a general increase in undrained shear strength for the entire profile highlights the potential cost savings resulting from extensive site investigation that incorporates evaluation the magnitude of strength anisotropy.

This article describes the development of a screening tool to estimate factors of safety and probabilities against slope failure for static and pseudo-static analyses, as well as seismically induced permanent displacements, for both onshore and offshore slopes. The screening tool combines geophysical and geotechnical data to rapidly assess and identify potentially unstable regions over large areas. The results allow a quick, objective, and logical rationale for selecting regions to perform more in-depth and detailed slope stability analyses, and/or areas to avoid when selecting locations or paths for future infrastructure development. We apply the screening tool to part of the continental slope offshore the Lofoten islands, Norway.

Since 2003, the Québec Government uses airborne lidar surveys to map landslide-prone areas. Hillshade maps greatly helped in obtaining a detailed inventory of old scars of retrogressive landslides in sensitive glacio-marine clays. This technique allowed a much better identification of scars than with conventional air photo interpretation. It has also made possible to distinguish different types of landslide, even when scars are very old and forested. Although only 66% of the area has been covered by lidar so far, about 3,500 scars of retrogressive landslides have already been counted in post-glacial Champlain, Laflamme and Goldthwait seas. The 50 largest scars identified so far were split into three groups of dimensions. The seven cases included in the first group have surface areas comprised between 6.5 and 20 km2, and most were triggered by earthquakes. The second group is characterized by landslide scars with a surface area between 2.5 and 5.3 km2. The only event of this group documented in historical records is the St. Alban landslide, with a surface area of 4.5 km2. The third group includes the last 23 cases that have surface areas comprised between 1.0 and 1.9 km2. The second largest event in recorded history is the 1896 spread of Grandes-Bergeronnes, with a surface area of 0.54 km2.

An essential part of landslide hazard and risk assessment is the estimate of the runout distance of the landslide masses. There is, however, little guidance available today on the estimation of the landslide runout in sensitive clays and no suitable model exists for predicting runout in sensitive clays. A new empirical model for the runout estimation is presented in this paper. The new model is based on empirical data, and is recommended for use in Norway until further research on analytical models becomes available. The recommended empirical procedure is based on the historical landslides in sensitive clays in Norway. The paper discusses the implementation of the proposed empirical models in a calculation tool called GeoSuite Toolbox as a part an ongoing R&D project GeoFuture II.

Sensitive clays are prone to various types of landslides. Among these are flow slides that are able to affect hectares of land. Moreover, debris from these flow slides has a high mobility with run out distance of hundreds of meters, even in relatively flat areas, are quite common. In the context of hazard mapping, mobility of the debris is also an important factors to consider. In this context, a parametric analysis using simplified geometries is undertaken in order to evaluate the run out characteristics of these flows (such has the length of the run out area and the lateral spread of the debris) as a function of the rheological parameters such as the yield stress and viscosity. In order to proceed with the parametric analysis, a newly developed 3D numerical model was used.

The Norwegian regional mapping of quick clay hazard zones uses a 1:15 inclined line from the foot of the slope to decide the maximal retrogression distance of a quick clay slide. When the hazard zones are subjected to a more detailed examination, the site investigations give a better picture of the location of the quick clay, which is decisive for what type of slide that may occur. Based on this, a method was developed through the research project “Natural hazards – infrastructure for floods and slides (NIFS)”. This has been applied to an existing hazard zone and compared with other existing methods to draw the depletion zone. The implementation shows the importance of first deciding the slide mechanism. If the extent of the quick clay is well mapped, the NIFS-method can be a useful tool to evaluate the retrogression distance, but it must be used with careful consideration of the geotechnical conditions.

The presence of the sensitive clays in Scandinavia and Canada creates quick clay landslide hazards. The ability to predict the likelihood of such landslides occurrence, as well as their outreach would reduce the damage to the infrastructures and loss of life. Recently, a simple experimental technique named a quickness test (Thakur V, Degago S, Geotech Lett 2:87–95, 2012) has been employed to investigate the susceptibility of the clay to create large retrogression landslides. In this paper, we applied Generalized Interpolation Material Point Method (a numerical method suitable for large displacement dynamic problems) to replicate the quickness test experiment. The primary goal of the presented simulations is to further validate the modelling technique and the constitutive model used. In particular, the computations suggest the importance of the strain rates for the prediction of the run-out distance of the remoulded sensitive clays.

This work presents a back-calculation of run-out of Byneset flow slide that took place in Norway in 2012. The flow slide involved about 350,000 m3 of sensitive clay. The Voellmy rheology is selected for this back-calculation. An attempt is made to verify whether this rheology can capture the actual run-out distance. The numerically obtained flow depth and flow velocity were evaluated. The results indicate that, this rheology was able to simulate the run-out distance for some range of values of the key input parameters, i.e. friction coefficient and turbulence factor. However, these parameters are not easy to relate to the well-established geotechnical parameters of sensitive clays e.g., liquidity index, remolded shear strength, quickness and viscosity. Therefore, more research is needed to understand the input parameters in terms of the geotechnical parameters. This is the key to explain how the Voellmy rheology works for sensitive clays.

Reliable prediction of landslide triggering threshold and landslide run-out distance is essential for hazard risk assessment. The paper focuses on studying slides in sensitive clays, which represent a major geohazard in many countries including Norway, Sweden and eastern Canada. Large deformation finite element (FE) analyses were performed using the Coupled Eulerian-Lagrangian (CEL) method in Abaqus, which allows for capturing of the full progressive failure mechanism (initiation, propagation and breakoff) involved in a sensitive clay slide. The 1984 slide in Vestfossen, Norway, was chosen as problem case of progressive failure in sensitive clay to be back-calculated by using the CEL FE-model. It is found that the failure mechanism predicted by the FE-analysis agrees reasonably well with the historical failure mode observed at Vestfossen. A parametric study has been performed on the remoulded shear strength as well as the rate of strain softening of the sensitive clay in order to evaluate their effects on the landslide run-out distance.

The landslide of 1908 in Notre-Dame-de-la-Salette, Québec, was the deadliest event occurring in sensitive clays of Eastern Canada, causing 33 deaths. Of these, 26 are associated with the tsunami generated impact of water and ice on the opposite bank. A LiDAR survey of the sector, and a geotechnical investigation were carried out respectively in 2009 and 2010 to characterize this landslide. Covering an area of 6.5 ha, the soil mass carried is estimated at nearly 1.2 million m3. The paper describes the event, reports the results of the investigation and discusses the tsunami caused by the debris of the landslide. The tsunami approach includes modeling both the kinematics of the slide and the wave.

Rambøll was responsible for the geotechnical design of a pedestrian road through an area with quick clay in Modum municipality Norway. This paper reviews the ground conditions and soil parameters, describes stability calculations and describes the design solution, including protection against erosion and the implementation of each construction phase. The design was done according to guidelines given by the Norwegian Public Roads Administration (NPRA) and The Norwegian Water Resources and Energy Directorate (NVE). The pedestrian road was considered to be an important project to improve the traffic safety for the pedestrians and cyclists in the area and the project was designed as a K1-project according to NVE guidelines. For K1-projects the design must ensure no reduction of the current factor of safety in any of the construction phases.To determine the soil parameters comprehensive field and laboratory tests were performed. Interpretation of CPTUs and laboratory data indicated a weak layer in the slope, approximately coinciding with the critical shear surface. 3D–effects of the slope were taken into account and the soil parameters were increased to the level where a safety factor of ca. 1,0 was attained. The pedestrian road was located half way up the slope and stability calculations were performed both for the slope above the road and for the slope between the road and the river Simoa below. The design solution included lightweight material for the new road filling to avoid increased loads. Movements in the slope and the pore pressure were monitored throughout the project.

This case study presents the results of geotechnical investigations and desktop studies carried out on a site characterized by the presence of a 55 m thick sensitive Champlain sea clay deposit, in the Province of Quebec (Canada). The main objective was to determine an iron ore stockpile configuration that meets the design factor of safety (FoS). The site is located in the St. Lawrence River lowlands. It is adjacent to the river and extends 200 m south of the shore line. For decades, the site has been subjected to temporary loadings which induced vertical settlement and small lateral deformations towards the River. This paper presents the inferred site stratigraphy, including geotechnical properties of the units encountered, presents the results of a 1D wave propagation analysis, stability analyses and preliminary deformation assessment with regard to site conditions. Field works performed involved seismic piezocone penetration tests (SCPTu) with the recording of pore water pressure, and geotechnical boreholes.

The Sokkelvik landslide and associated tsunami led to nine casualties and is considered one of the most devastating Norwegian quick clay landslide of the past century. In this paper we review the potential causes for this landslide based on an integrated study of eye witness testimony, swath bathymetry data and stability analyses. Results show that the Sokkelvik landslide was most likely triggered by intense rainfall and snowmelt, but that the main cause is associated to the load of an up to 7.5 m high embankment fill. This fill was placed at the shoreline for road construction 6 months before the landslide. Results presented herein show the importance of accounting for extreme rainfall events and groundwater flow when planning construction activity in sensitive near shore areas.

In geotechnical engineering, the presence of sensitive clays poses a major challenge. The landslides at Rissa in 1978, and more recently at the Skjeggestad bridge in Norway, are reminders of the potential devastating threats related to such soils. Norway has several sets of regulations to handle construction work in sensitive soils. This paper exemplifies the implementation of the regulations for a densely populated area on a quick clay area. The city area of Trondheim in central Norway has several zones of quick and sensitive clays which cause severe challenges for urban planning and development. One of these areas is the Gløshaugen – Bakklandet quick clay zone. The project partners consulted the company Multiconsult ASA to undertake the geotechnical evaluation of the area in the period 2012–2014. This work helped assessing areas where special geotechnical investigations and evaluations must be carried out to allow further development of infrastructure and erection of new buildings. The results from this work have been further developed into hazard maps, revealing several major deposits of quick clay within the quick clay zone. The results may be used as a basis for general planning by the local authorities, but are particularly helpful as a background for further development of the NTNU university campus at Gløshaugen.

The city of Saguenay, in the province of Quebec, is located in an area with a hilly topography and where sensitive marine clays are predominant. The slopes throughout the city are either former riverbanks or scarps of old large retrogressive landslides and their height can vary from a few meters to dozens of meters. Although the toe of these slopes is no longer subject to erosion, they can still be affected by shallow landslides. The spreading debris of such landslides is a threat to buildings located at the toe of slopes and thus a significant hazard for the safety of residents. This article presents the approach developed jointly by the Government of Quebec and the city of Saguenay to manage the risks associated with this hazard. Following the mapping of landslide prone areas, risk analyses allowed to bring out areas most exposed to shallow landslides. An annual inspection program was implemented in 2007 by the city of Saguenay and allowed on numerous occasions to detect early signs of landslides and take preventive action, in the form of stabilization work. This program has considerably reduced the risk to the population.

Quick clays involve considerable risks because small initial slips may evolve into large landslides involving the entire quick clay formation. Most large clay slides in Sweden, Norway and Canada have been in quick clay areas. It is therefore necessary to develop cost-effective methods for mapping the extent of quick clay formations and areas with probable quick clay. The most important areas are those with existing infrastructure and buildings as well as areas planned for exploitation. The methodology should be well-designed in proper steps with different levels of accuracy and include overview studies of topographical and geological methods, detailed geophysical and geotechnical ground investigations, as well as airborne geophysical methods with greater coverage but less accuracy. Three areas in southwest Sweden and one area in the northeast have been selected for this study. In these areas, locations with and without quick-clay have been identified by geotechnical sounding and sampling. Airborne electromagnetic (ATEM) measurements have been carried out in all four areas. The preliminary results from the first two areas, which are presented in this paper, show a reasonable correlation between the different methods used.

Identification of sediment types and in particular delineation of leached, possibly sensitive marine clays is of crucial importance for geotechnical design of infrastructure projects in Norway. Since leached clays normally have a lower salt content than intact marine clays, the electrical resistivity is consequently higher, and thus clay characterization may be based on data from high-resolution airborne electromagnetics (AEM) collected from helicopter. However, the resistivity difference between leached and unleached clays is small compared to the transition to bedrock and may furthermore vary locally. Therefore, indication of leached clays based on resistivity data has so far been done by manual interpretation. Here, we present a new procedure to calculate the likelihood of possible sensitive clays directly from AEM data. Geotechnical ground investigations are used to locally determine the expected resistivity of sensitive clay. The computation results are compared with well-known quick clay zones. The procedure is not intended as a simple solution to delineate quick clay, but to evaluate an area’s likelihood of sensitive clays that can be used as a cost-saving tool to efficiently place geotechnical investigations.

For several decades mapping of Quaternary geology has been the basis for quick-clay mapping in Norway. In this context it has been, and still is, of particular importance to define the distribution of fjord-marine deposits that commonly contain clays and silts where layers or pockets of quick clay may have developed. Quick clay may collapse under certain conditions and give rise to disastrous landsliding with severe consequences. It requires careful communication of Quaternary map information to get its full use for quick-clay mapping, landslide hazard assessment and other purposes. For this reason, there has been an increased focus in recent years at the Geological Survey of Norway (NGU) to improve existing web-based map services. This includes for example a National overview of the marine limit (ML) in Norway as the upper natural boundary for the occurrence of marine clays. In addition, a filtered version of Quaternary map information below ML has been added called clay-deposit susceptibility. This map service gives an overview over areas where clay deposits with some probability may be present even under other deposit types. The next step in the development of web services is to include information where the occurrence of fjord-marine clay deposits and quick clay are registered, for example, from drill-hole information. This is now made possible especially with the help of the newly established National Database of Ground investigations (NADAG) hosted by NGU.

Airborne transient electromagnetic (ATEM) data for mapping clay areas are acquired in four areas in Sweden. The resistivity models from the inversions of ATEM data are compared to the existing geotechnical, geological and ground geophysical data in one of the areas at Slumpån located in the Göta River valley. The ATEM models reveal information about layering and thickness of the sediments, the river depth and bedrock undulations. The estimated resistivities at the known locations of quick clays are within the range of 8–40 Ωm. The variation is dependent on the type of the surrounding sediments and the leaching process. The resistivity models have a limited resolution and must always be integrated with geotechnical and geological information for a confident and precise interpretation that leads to a realistic model. The method can be utilized as an effective tool prior to planning of any detailed and costly ground geotechnical investigations.

Mapping of the distribution and properties of marine clay is important in Norway due to numerous landslides in sensitive clay. The degree of leaching of salt in marine clay may be reflected by its electrical resistivity. However, the degree of sensitivity of the clay needs to be confirmed by geotechnical studies. Electrical resistivity and various electromagnetic (EM) methods are common geophysical methods to investigate the resistivity of an area. Helicopter EM surveys are helpful to investigate a large area in rather shorter time compared to ground EM or resistivity surveys. Frequency domain helicopter-borne EM (FHEM), electrical resistivity tomography (ERT) and seismic refraction were performed in 2014 at Byneset outside of Trondheim, Norway where a landslide occurred in 2012. Geotechnical surveys were performed in the region before, but mainly after the landslide. There was a good correlation between the results from the different surveys. FHEM interpretation revealed that unleached marine clay was covered by varying thickness of leached clay in the survey area. At some places, bedrock was very shallow and even exposed at the surface. FHEM is proven to be a very good tool to get an overview of the leached and unleached clay zones and to map 3D resistivity of the region.

Exploitation of the subsurface is becoming more frequent and the demand for knowledge about ground conditions is increasing. A vast amount of data from ground investigations such as geotechnical drilling, bedrock drilling and ground water wells exists in Norway. However, many of these are not easy to access as they are spread between multiple data owners and users. Following the development of The National Database for Ground investigations (NADAG) during the last years, the registration of geotechnical data has started. With increased accessibility of data, re-use will lead to considerable savings for the society. Importantly, the information will allow for better landslide hazard zonation. In addition, the effectiveness of emergency planning and response will improve with access to relevant, accurate and timely information about the local ground conditions. This may be crucial after landslide events with regard to the assessment of potential landslide expansion and the safe evacuation of people. NADAG aims to collect and make publically available data from ground investigation important for the society. The database contains various amounts of data, depending on availability – ranging from metadata (location, drill type, drill depth, company, date, report no., etc.) to full reports and raw data. NADAG will initially be populated by data from geotechnical investigations. A primary objective for NADAG is to distribute data from all types of ground investigations in Norway and to present the data coverage through a map-enabled web application.

The landslides at Rissa in 1978, and more recently at the Skjeggestad bridge in Norway, are devastating reminders of the potential threats related to quick clays. For a geotechnical engineering project it is hence important to determine if there is sensitive clay present and to clarify the extent of the quick clay deposit. Integration of geophysical and geotechnical methods has become more common in ground investigations nowadays, particularly in larger projects. In such integrated measurements, geotechnical engineers and geophysicists can cooperate, and by joint knowledge decide where geotechnical soundings, in situ tests and sampling should be located with optimal cost-efficiency. This paper describes how various investigation methods may be combined to achieve a successful strategy for detecting deposits of quick and sensitive clays. The methods presented herein include conventional soundings, CPTU and field vane test (FVT), supplemented by geophysical methods such as CPTU with resistivity measurements (R-CPTU), Electrical Resistivity Tomography (ERT) and Airborne Electromagnetic Measurements (AEM).

Risk and probabilistic analyses have now had enough applications that make them effective to use in practice. The approach provides more insight than deterministic analyses alone. They help reduce uncertainty and focus on safety and cost-effectiveness. The paper illustrates the use of reliability methods for the analysis of slopes in sensitive clays with examples of the calculation of probability of failure and run-out for the Finneidfjord and Rissa landslides in Norway. The input, model and results of the probabilistic slope analysis are described, including the uncertainties in the parameters, triggers and calculation model, as well as a brief review of the principles of the reliability approach. Reliability approaches do not remove uncertainty nor do they alleviate the need for judgment. They provide a way to quantify the uncertainties and to handle them consistently. Site investigations, laboratory test programs, limit equilibrium and deformation analyses, instrumentation, monitoring and engineering judgment are necessary inputs to the reliability approach. Landslide events, often unwittingly, are triggered or aggravated by human activity, such as change in topography (e.g. excavation or surcharge) and change in drainage conditions. Climate change can increase the frequency of landslide. The paper proposes that a probabilistic model in an event tree format should be included to ensure that all failure modes and the uncertainties have been covered and that slope failure mitigation measures are quickly available.

In several countries, including Norway, critical transport infrastructure is exposed to natural hazards such as floods, avalanches and landslides. Recent experiences show that these natural hazards have become more frequent primarily due to extreme weather events. These natural hazards have constitute a major threat to human life, surrounding manmade facilities and infrastructure and in Norway. Therefore, decision-making and design processes related to road and railroad construction in Norway rely heavily on an accurate assessment of these natural hazards. From 1900 to date, more than 1100 people have died in Norway due to natural hazards. About 115,000 people live in areas prone to flooding and landslides. According to the statistics, the accumulated cost of damage inflicted in the period 1980–2000 is around 1 billion Euros, and for the periode 2011–2014 the cost of the damage increased by approximately 750 million Euros.This paper presents some results from the NIFS research program to develop robust approaches to mitigate landslides and minimise the consequences on infrastructure such as roads, railways, housing etc. We present some aspects of real time warning of avalanches, landslides and flooding, application of technology for hazard mapping and early warning, the development of new national databases, and the national strategy for management of natural hazards. It is our belief that some of these technologies and methods adopted for the flood, landslide, and avalanche hazard management are equally applicable for the preparedness and the early warning systems required for landslides in sensitive clays.

The Government of Québec recently initiated the deployment of a vast groundwater pressures monitoring network in postglacial marine clays to document their variations in time and improve our understanding of the relationship between failure initiation and climate in clay slopes. This project aims at evaluating the impacts of climate change on clay-slope stability and how it can be integrated in landslide risk management to improve public safety. Hydrogeological data will be acquired at sites located throughout the Québec Province’s post-glacial clay deposits to create a public georeferenced index of typical hydrogeological conditions. The project goes beyond the characterization of groundwater pressures and their variations in clay slopes. Indeed, slope deformation will be measured at several sites. Also, two sites in flat terrain will be instrumented in order to evaluate mechanical properties of clay layers in simple 1-D conditions and groundwater recharge. The unsaturated clay crust in slopes susceptible to superficial landslides will be characterized and instrumented. The current lifetime of the monitoring project has been set to a period of 25 years.

Safety level of sensitive clay slopes can rigorously be established a priori for a given preconstruction state under all foreseeable conditions. However, construction activities may alter the in-situ conditions of the slope and possibly lead to a lower safety level or an increased level of hazard beyond an acceptable level. Therefore, it is crucial to have a thorough follow up during construction activities and implement hazard mitigation measures to ensure or maintain an acceptable safety level. This is illustrated in this paper using a real case involving construction of an infrastructure in highly sensitive clay slope in Rissa, Norway. The work emphasizes that as much as establishing factor of safety or probability of failure is important in design stage, one should also equally focus on trying to reassure the analysis/assessment done in design remains valid during and after construction.

The Swedish Geotechnical Institute (SGI) is assigned by the Swedish government to perform risk assessment for landslides in soft soil along priority watercourses as one part of the national climate adaptation funding. From the previously developed landslide risk mapping methodology applied for the Göta River Valley emerged a need for a less extensive methodology that could be implemented at significantly lower cost. Hence, the aim is to develop a sufficiently simplified methodology that could provide a basis for planners in municipalities and county administrative boards in their work with prioritization and preparation of adaptation measures. A criterion was also that such a methodology should ease the interpretation of the landslide risk maps, thereby increasing the societal relevance and usability of the results. The Norsälven River valley was used as a pilot area. The landslide risk analysis along the Norsälven River valley has resulted in a comprehensive overview of the landslide risk in the present and future climate, for built-up as well as undeveloped land and areas with vital infrastructure. The main implications of the climate change for the Norsälven River valley concerns the effect of increased erosion on the slope stability caused by increased water flow. The simplified methodology for landslide risk mapping that has been developed is applicable for landslide risk mapping along other river valleys. However, some modifications will be necessary due to site specific conditions.

The Norwegian Water Resources and Energy Directorate (NVE) is responsible for the national quick-clay hazard mapping in Norway, and for the presentation of the quick-clay hazard zones that are identified. The first quick-clay hazard mapping started in the beginning of 1980, in the southeast and middle parts of Norway, initiated by the big quick-clay landslide in Rissa, Norway, in 1978. Today, there are nearly 2000 mapped quick-clay hazard zones in Norway, and the mapping is still ongoing. In the last years (2012–2015), the methods for evaluating quick-clay hazard zones and the presentation of quick-clay areas have been developed through the multi-disciplinary research project Natural Hazards: Infrastructure for Floods and Slides (NIFS), which is a cooperation between the Norwegian National Rail Administration (NNRA), the Norwegian Public Roads Administration (NPRA) and the NVE. New and improved quick-clay hazard maps are one of the positive outcomes of this project. Increased accessibility to geotechnical data is important for the quick-clay mapping, which is improved through the new national database for ground investigations (NADAG). Another outcome of the NIFS project is the quick-clay areas mapped by the NPRA, which are shown together with the quick-clay hazard zones on the NVE website. The NIFS project has also developed new methods for run-out areas and shoreline evaluation, which from now on will be included in the mapping.