Proposed electricity generation developments: peat landslide hazard best practice guide

Hazard and risk assessment

5.1 Overview

Methodologies for geotechnical hazard and risk assessment are well covered in several existing publications ( e.g. Lambe and Whitman, 1979; Brunsden and Prior, 1984; Bromhead, 1986), while hazard and risk assessment for landslide investigations is thoroughly considered in Lee and Jones (2013). This revised section provides an overview of the principles of hazard and risk, with specific reference to peat landslides, and is informed by these documents and by use of the previous guidance by statutory bodies, developers and assessors.

It should be noted that examples provided in the following sections are illustrative only and should not be taken as prescriptive or used as a substitute for a developer's preferred methodology. In all cases, risk assessment methodologies should be clearly explained and incorporate consideration of the likelihood of instability and the consequences should it occur.

5.2 Hazard and risk assessment

There is no universal, agreed definition of hazard and risk that can be applied in the context of peat landslides. However, there is general acceptance in the literature that risk is 'the potential for adverse consequences, loss, harm or detriment'. Peat landslides can be considered a type of geohazard, i.e. 'a geological process that in particular circumstances could lead to loss or harm'.

Risk can be expressed as the product of the probability of a [peat] landslide event and its adverse consequences (after Lee and Jones, 2013; see p104), i.e.:

The definition as written above is particular to peat landslides, but has the same meaning as previously presented risk definitions in Clayton (2001) (Eq. 2), and in Winter et al. (2008) (Eq. 3):

The probability of a peat landslide reflects the combined influence of preconditions, preparatory factors (see Section 2.2), and triggering factors (see Section 2.3), or collectively 'controls', on the stability of a peat deposit. The updated version of this guidance considers peat landslide hazard as a consequence of both natural controls and man-made controls.

The addition of man-made controls (such as construction activities, alterations to peat drainage) reflects the potential destabilising effects of human activity on peatlands, and evidence from well publicised peat landslide events that human activity may exert a significant control on peat stability ( e.g. Lindsay and Bragg, 2004; Dykes and Warburton, 2007a). The revised definition of risk reflects increased emphasis in this issue of the guidance on considering all potential adverse consequences.

Adverse consequences may include accidents, loss of life, environmental impacts or damage to site infrastructure and associated financial losses. The potential for adverse consequences reflects the exposure to peat landslide hazard of elements at risk within a specific area. For example, infrastructure may always be exposed to the effects of a peat landslide, but site traffic may only occasionally be exposed.

There are various levels of risk assessment, ranging between:

• High-level qualitative assessments where the objective is to develop an approximate estimate of the risks, particularly in relative terms ( e.g. low, medium and high levels of risk);
• Detailed quantitative risk assessments (or QRA) where the objective is to generate more precise measures of the risks ( e.g. expressing risk as a specific magnitude of loss).

As part of the EIA submission, it is expected that a PLHRA provides sufficient estimate of risks to enable infrastructure layout ( e.g. turbines, hard standings, compounds, access tracks) to avoid areas of medium or high risk, while also making full and detailed recommendations for mitigation of low and medium risks where exposure remains. In the case of peat landslides, there may also be a desire to compare economic and environmental risks associated with a particular development proposal.

The general principles of risk assessment are similar, regardless of the elements at risk ( e.g. people, peat resource, development programme). The next section provides guidelines on approaches to risk assessment in relation to peat landslides.

5.3 Assessing the likelihood of a peat landslide

5.3.1 Probabilistic or factor based approaches

There are a number of approaches to estimating the likelihood or probability of a future landslide in a given area (for a full account with worked examples, see Lee and Jones, 2013):

a) Use the historical frequency of landslides in the area to provide an indication as to future annual probability, i.e.:

For example, if 2 landslides have occurred in 100 years in a particular area of peatland, the annual probability is:

A 2% annual probability of occurrence also means a 98% probability of non-occurrence in any one year.

The probability of a peat landslide event occurring in a 25 year period (a typical wind farm design life) is equivalent to (after Lee and Jones, 2013):

The approach is relatively easy to apply where there is a historical record of landslides. However, this approach assumes that conditions in the future correspond to conditions in the past, which is not necessarily the case, for example, the construction of a wind farm may elevate likelihood through alterations to natural drainage pathways.

b) Use the probability of a landslide triggering event as an indicator of the probability of a landslide, assuming other preparatory factors indicate peat slopes to be of marginal stability.

In the case of peatlands, a number of publications cite triggering threshold rainfall intensities in association with recorded peat instability and in association with specific ground conditions (see Evans and Warburton, 2007; Dykes and Warburton, 2007a; GSI, 2006). Other natural triggers may be used ( e.g. snowmelt), or man-made triggers can be considered ( e.g. slope cutting or loading).

c) Estimate probability through expert judgement, whereby general principles are used to assign probabilities to landslide scenarios.

Such approaches may use ranking systems that relate ground conditions to the probability of landslide occurrence, e.g. the presence or absence of instability features at the site, or combinations of scored 'hazard factors' ( e.g. slope, peat depth, orientation of slope drainage) whereby higher scores indicate higher probability of future peat landslides.

Where expert judgement is used, judgements should be transparent through full documentation of sources of evidence, and the logic behind any factoring or scoring approach should be clearly detailed.

d) Estimate probability through stability analysis, i.e. by providing a quantitative measure of slope stability incorporating consideration of slope form (slope angle), materials (shear strength), loadings (overburden) and transient parameters ( e.g. pore pressure).

The results of stability analysis are generally presented in terms of a Factor of Safety (F), where:

The probability of a peat landslide occurring should be based upon the probability of a Factor of Safety being less than 1 over the period of concern ( e.g. 25 years).

A full explanation of stability analysis is beyond the scope of this guidance, however, where it has been applied ( e.g. Carling, 1986; Warburton et al. 2004; Long and Jennings, 2006; Dykes and Warburton, 2008), the infinite slope model has provided the most informative results. A short summary is provided below.

5.3.2 Stability analysis

The likelihood of a particular slope or hillside failing can be expressed as a Factor of Safety. For any potential failure surface, there is a balance between the weight of the potential landslide (driving force or shear force) and the inherent strength of the soil or rock within the hillside (shear resistance) ( Figure 5.1). Provided the available shear resistance is greater than the shear force then the Factor of Safety will be greater than 1.0 and the slope will remain stable. If the Factor of Safety reduces to less than 1.0 through a change in ground conditions e.g. through the application of a trigger, the slope will fail.

The shear force is mostly a component of the weight of the rock/soil/peat making up the potential landslide mass. The shear resistance is provided by the frictional and tensile strength of the peat, soil or rock, and the normal effective stress and influence of groundwater increasing the normal load (weight) of the sliding mass (Warburton et al., 2004; Dykes and Warburton, 2007b; Dykes and Warburton, 2008). The field sampling methods and laboratory tests recommended in the previous chapter provide the means of quantifying these controlling parameters in a ground model of the development site.

Where stability analysis has been applied to peat slopes ( e.g. Carling, 1986; Warburton et al. 2004; Dykes and Warburton, 2008), the infinite slope model has provided the most informative results. The infinite slope model assumes a planar translational failure, where the shear surface is parallel to the ground surface, and the length of the slope is large in comparison to the failure depth (hence 'shallow' failure). The nature of detachment of peat landslides is most frequently by a translational mechanism, and since this is the failure type modelled by infinite slope analysis, it is the most appropriate analytical method.

The stability of a slope can be assessed by calculating the factor of safety F, which is the ratio of the sum of resisting forces (shear strength) and the sum of the destabilising forces (shear stress):

where c′ is the effective cohesion, γ is the bulk unit weight of saturated peat, γ _w is the unit weight of water, m is the height of the water table as a fraction of the peat depth, z is the peat depth in the direction of normal stress, β is the angle of the slope to the horizontal and ϕ′ is the effective angle of internal friction.

Values of F < 1 indicate a slope would have undergone failure under the conditions modelled; values of F > 1 suggest conditions of stability.

The infinite slope model can be modified to allow use of 'slices' in the slope (Craig, 1997). These slices allow sections of the slope with differing characteristics, such as peat depth or slope angle, to be treated individually. By considering the length of the slices, a residual mobilising force from one slice, if unstable, can be brought to bear on the slice below and taken into account in the stability analysis of the lower slice. Slices are not modelled as providing restraining forces to the slices below, as this is highly unlikely to happen in practice.

Sufficient slope stability analysis should be undertaken to represent the range of material, topographic and hydrological conditions at the development site. Spatial variability in Factor of Safety can then be used as a key input into hazard zoning or for calibration of other approaches to estimating landslide probability (such as factor based approaches).

5.3.3 Summary

Ultimately, under the assumption that peat landslide probability is spatially variable over a proposed development area, and assuming that this spatial variability reflects site conditions, probability ranges can be assigned to different areas of the site (see Table 5.1). When compared with the potential for adverse consequences in the same polygon, a risk value can be determined for that specific location in the proposed development area.

The table below provides an illustration of how qualitative descriptions of likelihood might relate to the probability of a landslide occurring. The developer should not consider these examples as prescriptive and should determine any likelihood scale (if used) according to their own understanding of the site conditions and with respect to the hazard and risk assessment methodologies they have chosen.

Table 5.1 Peat landslide probability ranges for the lifetime of a proposed development

Scale	Likelihood	Probability of occurrence
5	Almost certain	> 1 in 3
4	Probable	1 in 10 - 1 in 3
3	Likely	1 in 10 2 - 1 in 10
2	Unlikely	1 in 10 7 - 1 in 10 2
1	Negligible	< 1 in 10 7

5.4 Assessing adverse consequences

Potential adverse consequences, in the event that a peat landslide does occur, should be estimated. The intention should be to represent consequences as a range that can be applied to specific areas of the development site. For example, the consequences of a landslide occurring for a watercourse may depend on how far the landslide is from it, or on the importance of a watercourse from a habitat perspective, e.g. it may be designated as a Special Area of Conservation ( SAC).

Consequences could include but are not limited to the following elements:

• The potential for harm to life during construction;
• The potential economic costs associated with lost infrastructure, or delay in programme;
• The potential for reputational loss associated with occurrence of a peat landslide in association with construction activities;
• The potential for permanent, irreparable damage to the peat resource (both carbon stock and habitat) associated with mobilisation (and ultimately loss) of peat in a landslide; and
• The potential for ecological damage to watercourses subject to inundation by peat debris.

A magnitude of adverse consequence should be attached to each element for each area to which a peat landslide probability has been assigned (see Table 5.2).

Table 5.2 Degree of adverse consequence for elements exposed to peat landslide hazard (example shown here relates to financial cost)

Scale	Adverse consequences	Impact as % damage to (or loss of) receptor
5	Extremely high	> 100% of asset ( e.g. infrastructure or habitat)
4	Very high	10% - 100%
3	High	4% - 10%
2	Low	1% - 4%
1	Very Low	< 1% of asset ( e.g. infrastructure or habitat)

Consequence analysis is a key element of risk assessment, since without it only landslide potential or hazard is assessed. When determining receptors at risk, care should be taken to identify designated environmental features both within the application boundary and within a reasonable distance of it (since the effects of peat landslide debris may be felt off-site if debris is incorporated within watercourses).

5.5 Determining risk

Once all areas within the site have been assigned a peat landslide probability and a degree of adverse consequence, a risk level can be estimated for each area, and peat landslide risk maps prepared for the development site. Table 5.3 shows how the qualitative descriptors for likelihood and adverse consequences can be simply combined to produce risk levels.

Risk maps can be produced for individual elements at risk ( e.g. environment, infrastructure) or a summary risk map which sums the risks associated with all exposed elements within the development area can be presented.

The need for further investigation or specification of mitigation measures should be a function of the risk level present on the site (see Table 5.4).

Table 5.3 Indicative risk levels

	Adverse consequence
	Extremely High	High	Moderate	Low	Very Low
Peat landslide probability or likelihood
Almost certain	High	High	Moderate	Moderate	Low
Probable	High	Moderate	Moderate	Low	Negligible
Likely	Moderate	Moderate	Low	Low	Negligible
Unlikely	Low	Low	Low	Negligible	Negligible
Negligible	Low	Negligible	Negligible	Negligible	Negligible

Where the risk level for a zone is medium or high, avoidance or specification of mitigation measures would normally be the only means by which project infrastructure could be considered acceptable within that zone at the proposed development site.

Table 5.4 Risk ranking and suggested actions

Risk Level	Action suggested for each zone
High	Avoid project development at these locations
Medium	Project should not proceed unless risk can be avoided or mitigated at these locations, without significant environmental impact, in order to reduce risk ranking to low or negligible
Low	Project may proceed pending further investigation to refine assessment and mitigate hazard through relocation or re-design at these locations
Negligible	Project should proceed with monitoring and mitigation of peat landslide hazards at these locations as appropriate

It is accepted that for some locations, further detailed ground investigation may be required to better quantify risks. If these locations are of concern, the developer should consider offering assurances through specification of these additional works in conditions of consent.

Possible mitigation measures in relation to peat instability are considered below.

5.6 Mitigation

The extent of mitigation required will depend upon the scale of the project, the nature of the risk and estimated risk level. A combination of options may be required to reduce the risks to an acceptable level for a given scheme.

5.6.1 Avoidance

Areas exhibiting medium or high risk levels associated with peat landslides should be avoided, for example by relocating infrastructure within the development area or by relocating to an alternate site. Where avoidance is not possible, the proposed design should be modified to incorporate engineering measures to reduce or eliminate the assessed risk.

5.6.2 Engineering mitigation measures to minimise landslide occurrence

Many of the site specific ( e.g. peat depth, slope angle) and site independent variables ( e.g. weather) that contribute to the incidence of natural peat landslides are beyond engineering control without significant damage to the peat itself. However, a number of engineering measures exist to minimise the risks associated with potential triggers (such as short term peaks in hydrogeological activity).

5.6.2.1 Installation of drainage measures

Installation of targeted drainage measures would aim to isolate areas of susceptible peat from upslope water supply, re-routing surface (flushes/gullies) and subsurface (pipes) drainage around critical areas. Surface water drainage plans should be considered as a useful way of accounting for modified flows created by construction, which in turn may affect peat stability, pollution and wildlife interests. Drainage measures need to be carefully planned to minimise any negative impacts.

5.6.2.2 Construction management

Site specific procedures aimed at minimising construction-induced peat landslide hazards should be identified and implemented and followed rigorously by site construction personnel. These may include work method statements subject to an environmental check to monitor compliance. These checklists should incorporate a weather forecast to avoid peat working during heavy rain and to allow environmental mitigation measures to be put in place where construction work is on-going. Weather forecasts can be obtained using data available from numerous web-sites or provided at a cost by commercial organisations or the Met Office. Particular care should be taken in relation to storage of excavated peat deposits on site, with loading of intact peat by excavated deposits avoided wherever possible.

Further guidance in relation to the construction of tracks on peatlands, and the management of peat on construction sites is provided by SNH and SEPA (Scottish Natural Heritage, 2015b; Scottish Renewables et al, 2015).

5.6.3 Engineering mitigation measures to control landslide impacts

A number of engineering measures are available for reducing the impacts (or exposure) associated with peat landslide hazards.

5.6.3.1 Catch wall fences

Where the potential for peat landslides has been identified, catch fences positioned downslope of the suspected or known landslide prone area can slow or halt runout (Tobin, 2003). Catch fences should be engineered into the peat substrate. Fencing may require periodic inspection for removal of debris.

5.6.3.2 Catch ditches

Similarly, ditches may also slow or halt runout, although it is preferable that they are cut in non-peat material. Simple earthwork ditches can form a useful low-cost defence. Paired ditches and fences have been observed (Tobin, 2003) to slow peat landslide runout at failure sites.

5.6.4 Monitoring and review

Risks identified in the PLHRA should be included in either a specific peat landslide hazard management plan or be included within the Geotechnical Risk Register for the site. These documents should be prepared by the developer or contractor undertaking the assessment, and passed on post-consent to the design and build contractor at the earliest opportunity. The implementation of measures in these documents (and update, if required) is as critical to safe site development as is their preparation.

It should be noted that factors that affect the likelihood of peat landslides and their consequences may change with time. Thus, ongoing review of the peat hazard management plan is essential. Design of stabilisation measures may be reviewed and risks may be reassessed during construction as the process of construction yields further data.

5.7 Post construction and restoration works

The PLHRA must also consider the risks (residual and new) following construction, and on completion of restoration works for the site, particularly where proposing to re-use excavated peat (Scottish Renewables and SEPA, 2012). Restoration proposals should aim to restore the water table of the peatland to ensure that the peatland becomes active again and therefore stores carbon. Otherwise, potentially significant changes to the hydrology of the peat bog may result in irreversible changes to the physical characteristics and structure of the peat that could both increase the likelihood of peat landslides and lead to long term degradation of the peat resource. Further Reading

5.8 Further Reading

This document has provided a brief overview of peat landslide hazards, the means by which they may be identified, the risks associated with their occurrence, and the measures that might be considered to mitigate these risks. A considerable range of published material, much of which is referred to herein, is available that may be of use to developers in undertaking the works described in this document.

The form, mechanisms and causes of peat landslides are considered in detail in a paper by Dykes and Warburton (2007a), and in two books that consider wider issues of peat degradation (Evans and Warburton, 2007; Martini et al. 2006).

The geotechnical properties of peat are well described in a chapter on organic soils by Bell (1999), with detailed discussion of peat physical and chemical properties provided by Hobbs (1986).

Methodologies for landslide investigation and management are provided in Department of Environment (1996), while landslide hazard and risk assessment is considered in detail in Lee and Jones (2013).

5.9 Acknowledgements

We gratefully acknowledge the input of CH2M Earth Engineering staff in the compilation of this document and the guidance given by Dr Jeff Warburton of Durham University and Dr Andy Mills of AM Geomorphology Ltd.

Thanks are given to nPower Renewables Ltd. For their kind permission to use the diagrams and data from the geotechnical input to Farr Wind farm.

The reissued version of this guidance has benefitted from discussions held during and subsequent to the "Reinforced Water: Engineering and Environmental Considerations in Construction over Peat" workshop convened by the Geological Society of London in Edinburgh in 2008.