Keywords

Introduction

SafeLand is a large-scale integrating collaborative research project funded by the Seventh Framework Programme for research and technological development (FP7) of the European Commission. Thematically the project belongs to Cooperation Theme 6 Environment (including climate change), Sub-Activity 6.1.3 Natural Hazards. The project team composed of 27 institutions from 13 European countries is coordinated by Norwegian Geotechnical Institute (NGI).

SafeLand will develop generic quantitative risk assessment procedures as well as tools and strategies for landslides management at local, regional, European and societal scales.

Within the general framework of the Project, the objectives of Work Package 5.1 are to provide:

  • A Compendium of tested and innovative structural and non-structural (including risk-transfer) mitigation measures for different landslide types (Deliverable 5.1);

  • A web-based “toolbox” of risk mitigation strategies and guidelines for choosing the most appropriate risk management strategy, based on technology, experience and expert judgment in Europe and abroad (Deliverable 5.2).

Besides feeding into the web-based toolbox, the Compendium is intended also as a stand-alone resource providing technical guidance on mitigation measures to a wide variety of end users. It has been compiled by the author and colleagues at Studio Geotecnico Italiano, with contributions from ICG of Norway, also responsible for quality assurance, AMRA and the University of Salerno from Italy, Aristotle’s University of Thessaloniki, Greece, Zurich Technical University and the University of Lausanne, Switzerland and the Geological Institutes of Slovenia and Romania. It is currently under validation and will be made available in 2012.

Continuous technological progress and innovation make it impossible to provide an exhaustive and detailed list of mitigation measures. Each of the techniques or approaches described in the compendium could have many variations, reflecting differences resulting for example from specific conditions which vary form place to place; technological development; commercial interests to differentiate products to overcome patents and copyright; different or changing legislation. Apparent variations may result also from the use of different terminology to describe substantially the same measure.

While every effort has been made to provide a comprehensive and balanced compendium, inevitably readers will note omissions and, possibly, apparent repetition.

In drafting the compendium, particular emphasis has been placed on providing a rational framework applicable to all the measures listed in the compendium and to any other specific measure that may be developed in the future. In the context of the SafeLand Project and in light of the general consensus on a risk based approach to landslide management, the classification of mitigation measures has been related to the term of the “risk equation” which is specifically addressed by each mitigation measure.

Classification of Mitigation Measures

It is widely accepted and is the backbone of the SafeLand Project that the management of landslides and engineered slopes involve some form of risk assessment and risk management (Ambrozic et al. 2009).

Figure 1 summarizes the framework for landslide risk management; it is widely used internationally and has been adopted as the reference framework in the “Guidelines for landslide susceptibility, hazard and risk zoning for land use planning” published by Fell et al. (2008).

Fig. 1
figure 00871

Framework for landslide risk management (After Fell et al. 2008)

Full size image

The evaluation, implementation and control of mitigation measures fall within this framework and in fact complete and complement the risk analysis and risk assessment stages of the process. It is therefore useful to relate the classification of mitigation measures to the same principles and criteria used in the rest of the process, using the internationally accepted definitions (Fell et al. 2008), which have also been adopted for the SafeLand Project and are repeated here for the avoidance of doubt: Hazard (H i ); Vulnerability (V i ) and Elements at risk (E) and Total Risk (R ti ).

The Total Risk Rti due to a particular (ith) phenomenon within a specified period of time and within a given area can be expressed as:

$$ {{R}_{{ti}}} = (E)\cdot ({{H}_i}\cdot {{V}_i}) $$

The Total Risk Rt from all (N) possible landslide phenomena within a specified period of time and within a given area is the sum of the risk posed by all the specific (ith) phenomena that impinge on the area of interest, subject to considerations of conditional probabilities of occurrence and to “domino chains”, i.e. the progressive triggering of distinct phenomena in a linked sequence of cause and effect (e.g. large landslide → natural dam → overtopping → debris flow etc.).

$$ {{R}_t} = \sum\limits_{{i = 0}}^N {({{E}_i})\cdot ({{H}_i}\cdot {{V}_i})} $$

It is evident that the Total Risk can be mitigated by reducing (see for example Canuti and Casagli 1994):

  • The Hazard (i.e. the probability of occurrence of one or more phenomena);

  • The Vulnerability (i.e. the the degree of loss to the elements at risk for a given hazard);

  • The Elements at risk (i.e. their number and/or specific value).

This represents a useful basis for classifying mitigation measures, because it provides a direct link with quantitative risk assessment and it highlights where the benefits of the mitigation measure being considered are accrued.

Other classifications of mitigation measures have been proposed, based on similar concepts but expressed in different terms. For example, Evangelista et al. (2008) distinguish between:

  • Stabilization: measures which increase the “margin of safety” of the slope or that intercept the run out (structural measures);

  • Restrictions on the use of the element at risk: permanently or temporarily;

  • Restrictions on land usage: through land-use planning tools, to limit the presence of elements at risk in the area threatened by the landslide (non-structural measures);

  • Actions by the Civil Protection authorities: which allow to remove from the area threatened by the landslide within a suitably short reaction time most valuable elements at risk, including as a minimum human life (emergency plans).

Similarly, Ambrozic et al. (2009) identify the following possible strategies for risk management:

  • Avoidance: can be implemented at the land-use planning stage for proposed development and/or to relocate existing facilities, if possible;

  • Tolerance: can be implemented if the risk level is deemed to be sufficiently low such that direct or indirect costs associated with other strategies cannot be warranted. Possible actions include “do nothing” or risk reallocation through private insurance or public intervention such as declaration of a “state of emergency” and the awarding of special funding and compensation to victims;

  • Monitoring/warning: can be implemented when landslide hazards affect large territories or when dealing with massive potential landslides. It provides additional information to enhance risk assessment and allows the implementation of warning systems for the temporary evacuation of the population at risk;

  • Stabilization: requires the implementation of engineering works to reduce the probability of occurrence of landslides;

  • Control works: requires the implementation of engineering works to protect/reinforce/isolate the elements at risk from the influence of landsliding.

Ambrozic et al. (2009) also refer more generally to:

  • Measures to reduce the hazard (through reducing the probability of triggering through stabilization and/or by reducing subsequent ground movement through barriers or containment);

  • Measures to reduce the vulnerability (i.e. reducing the consequences of failure).

This last statement exemplifies some of the difficulties that arise in classifying mitigation measures. In particular:

  • Although it may be justified in some respects to classify barriers and containment as hazard reducing measures, in the context of area wide risk management they might be better classified as measures to reduce the exposure of the elements they protect;

  • Avoidance may be as effective at reducing the consequences of failure as reductions in vulnerability, so inferring an exclusive association between reducing vulnerability and reducing the consequences of failure can be misleading.

Some of these difficulties derive from the definition of “vulnerability”, which Ambrozic et al. (2009) extend to include not only the damage functions with respect to ground movement (vulnerability s.s.), but also the number of the vulnerable elements potentially affected by a landslide and the probability that they will intersect the landslide ground movement.

Warning/alarm systems associated with plans for emergency evacuation or safe sheltering are often classified as measures to reduce vulnerability. However, keeping to the distinct definitions of “vulnerability” and “elements at risk”, these systems are best classified as measures to reduce (temporarily and selectively) the elements at risk, rather than their vulnerability.

Other somehow related, widely used, classifications of stabilization measures include distinctions between:

  • “Active” and “passive” stabilization measures (Picarelli and Urcioli 2006; Evangelista et al. 2008), in relation to whether the mitigation measures “actively” pursue an improvement s.s. of the stability of slope, or they “passively” intercept the run out when movement actually occurs, protecting the elements at risk.

  • “Hard” and “soft” stabilization measures (Parry et al. 2003a, b), where “hard” is normally used to describe structural techniques that are visually obvious, while “soft” is normally used to describe techniques that are visually less intrusive and which improve the strength or other properties of the ground, such as its drainage capability. The terms “hard” and “soft” can also be used in relation to the relative stiffness of the stabilization works and the surrounding soil, which results in the overall behaviour of the stabilized slope being modelled as an equivalent continuum or as distinct materials. “Hard” and “soft” can also be used in direct analogy with the terms “structural” and “non structural”, with the same meaning of hardware and software, depending on whether the mitigation measure addresses tangible, material or intangible, “immaterial” aspects of the risk.

  • “Preventive” and “remedial” stabilization measures (Parry et al. 2003a, b), relating to their relevance to different stages of movement (see Leroueil 2001).

Criteria for Selection

The selection of the most appropriate mitigation measures to be adopted in specific situations must take into account the following aspects:

  • Factors which determine the hazard, in terms of the type, rate, depth and the probability of occurrence of the movement or landslide, such as, for example the physical characteristics of the geosystem which can determine the occurrence of movement or landslides, including:

    • The stratigraphy and the mechanical characteristics of the materials,

    • The hydrological (surface water) and the hydrogeological (groundwater) regime,

    • The morphology of the area, and

    • The actual or potential causative processes affecting the geosystem;

  • Factors which affect the nature and quantification of risk for a given hazard, such as the presence and vulnerability of elements at risk, both in the potentially unstable area and in areas which may be affected by the run-out;

  • Factors which affect the actual feasibility of specific mitigation measures, such as:

    • The phase and rate of movement at the time of implementation,

    • The morphology of the area in relation to accessibility and safety of workers and the public,

    • Environmental constraints, such as the impact on the archaeological, historical and visual/landscape value of the locale,

    • Pre-existing structures and infrastructure that may be affected, directly or indirectly, and

    • Capital and operating cost, including maintenance.

Measures to Reduce Hazard

Mitigation measures which aim to reduce the hazard must reduce the probability of triggering of the landslide(s) which the specific measure is intended to address. Since triggering is caused by a decrease in shear strength Στr and/or an increase in driving shear stress Στd, mitigation measures which aim to reduce the hazard of landslides occurring must act in the system in the opposite direction, by increasing the resisting forces; and/or decreasing the driving forces.

Whilst it is clearly recognized that landslides are almost always the result of a combination of processes, in the Compendium hazard mitigation measures are subdivided in relation to the physical processes involved, as summarized in Table 1.

Table 1 Hazard mitigation measures (Adapted from Popescu and Sasahara 2009)
Full size table

Each class of hazard mitigation measures is briefly described and discussed in the main text of the Compendium and in greater detail in fact sheets attached thereto. Each fact sheet includes a brief description, guidance on design, schematic details, practical examples and references, as well as subjective (provisional) ratings of the applicability of the specific mitigation described in relation to the descriptors used for classifying landslides. Figure 2 shows sample pages from a typical fact sheet.

Fig. 2
figure 00872

Selected pages from a typical fact sheet of hazard mitigation measure

Full size image

The Compendium also includes a review of triggering and hazard mitigation measures investigated by physical models, including interaction of a row of piles with an unstable soil layer, reinforced soil retaining wall under dynamic loading, reinforeced soil structures, rainfall induced landslides, thawing of ice in rock joints and rock anchors, stabilisation effects of plant roots, soil nailing and monitoring the integrity of ground anchorages.

Measures to Reduce Vulnerability

Mitigation measures which aim to reduce vulnerability consist of “passive” solutions which are not intended to prevent the triggering of the landslide but to reduce the resulting degree of loss. The measures described in the Compendium can be subdivided in two main categories, as detailed in Table 2.

Table 2 Vulnerability mitigation measures
Full size table

Measures to Reduce the Elements at Risk

Mitigation measures which aim to reduce the elements at risk are particularly cost effective, especially when the number of elements at risk is small in relation to the extent of the landslide. The measures described in the Compendium can be subdivided in two main categories, as detailed in Table 3. A fact sheets on mitigation through reduction of exposed population through early warning systems is included in the Compendium. Further details can be found in the relevant deliverables of the SafeLand project on this subject.

Table 3 Measures to reduce the elements at risk
Full size table

Measures to Share Residual Risk

Among the possible strategies to manage landslide risk, techniques can be identified to increase the tolerance towards the residual risk that typically characterizes real situations even after implementing all other (technically and economically) possible mitigation measures. Of particular interest are risk sharing arrangements, which can be either voluntary or enforced. The two main mechanisms for this are:

  1. (a)

    Voluntary or compulsory insurance;

  2. (b)

    Compulsory systems based on taxes and public intervention in case of need.

The role and mechanism of insurance (private or public) is of particular interest and is discussed in detail in the Compendium, together with an overview of the natural hazard insurance system in Switzerland.

Reference in the Compendium to insurance and reinsurance companies can be taken to refer equally to private and public institutions, depending on local practice. Where Public Authorities replace private insurance companies, they face the same issues and have the same overall objective of loss reduction and efficiency.

Select National Experience

The Compendium includes an Annex which details National experience with landslide risk management in Romania, Slovenia and Switzerland.

The contribution on experience in Romania includes a general discussion on the national policy and practice in disaster management, risk assessment and systems for post-disaster impact assessment, which provides the backdrop for the presentation of national practice on landslide hazard mitigation.

The contribution on experience in Slovenia focuses on landslide hazard mapping, databases and the use of neural networks as tools underlying current practice in landslide risk management in the country.

Finally, the contribution on experience in Switzerland describes a number of case histories, showing how different mitigation measures are used to suit the specific conditions of each site.