An overview of peat landslides
2.1 Peat landslides
Peat instability is manifest in a number of ways, all of which can be observed on site or remotely from aerial photography, where photo resolution allows:
- Minor instability: localised, small scale development of tension cracks, tears in the acrotelm (upper vegetation mat), compression ridges, or bulges and thrusts; these features may be warning signs of larger scale major instability (such as landsliding) or may simply represent a long term response of the hillslope to drainage and gravity, i.e. creep;
- Major instability: comprising various forms of peat landslide, ranging from collapse and outflow of peat filled drainage lines/gullies (occupying a few-10s cubic metres), to medium scale peaty-debris slides (10s to 100s cubic metres) to large scale peat slides and bog bursts (1000s to 100,000s cubic metres) (Dykes and Warburton, 2007a).
Dykes and Warburton (2007a) provide a helpful overview and classification of the various types of 'major instability' referred to above. Forms of 'minor instability' are considered further in Section 2.2.
Two broad groups of peat landslide have previously been reported ( Plates 2.1 and 2.2). The term 'peat slide' is generally used to describe slab-like shallow translational failures (Hutchinson, 1988) with a shear failure mechanism operating within a discrete shear plane at the peat-substrate interface, below this interface, or more rarely within the peat body (Warburton et al., 2004). The peat surface may break up into large rafts and smaller blocks which are transported downslope mainly by sliding. Rapid remoulding during transport may lead to the generation of organic slurry in which blocks are transported. Peat slides correspond in appearance and mechanism to translational landslides ( DoE, 1996) and tend to occur in shallow peat (up to 2.0m) on moderate slopes (5 to 15°). A great majority of recorded peat landslides in Scotland, England and Wales are of the peat slide type.
The term 'bog burst' has been used to describe particularly fluid failures involving rupture of the peat blanket surface or margin due to subsurface creep or swelling, with liquefied basal material expelled through surface tears followed by settlement of the overlying mass (Hemingway and Sledge, 1941-46; Bowes, 1960). They are characterised by pear shaped areas of disturbed (often sunken) blanket bog, arranged in concentric tears and rafts, with little substrate revealed, and without necessarily a clear scar margin. Downslope of the area of subsidence, there is usually a block and slurry runout zone, similar in appearance to that associated with peat slides. Bog bursts correspond in appearance and mechanism to spreading failures ( DoE, 1996) and tend to occur in deeper peat (greater than 1.5m) on shallow slopes (2 to 10°) where deeper peat deposits are more likely to be found (Mills, 2002). Reports of bog burst failures are generally most common in the Republic of Ireland and Northern Ireland.
Both peat slides and bog bursts may have extensive runout, particularly where debris is incorporated in watercourses. The low density of peat debris means that it may be transported kilometres downstream with potential impacts on in-stream engineering structures and aquatic habitats (McCahon et al., 1987).
There is considerable natural variability in movement types and complex failures may result where the geotechnical properties of the peat vary. Hence there is some degree of overlap in processes and mechanisms between different landslide types. This variability is captured in the formal classification of peat landslides by Dykes and Warburton (2007a), based on a comprehensive database of examples collated from the literature and field studies. The classes of peat landslide reflect: the type of peat deposit (raised bog, blanket bog, or fen bog), location of the failure shear surface or zone (within the peat, at the peat-substrate interface, or below), indicative failure volumes, estimated velocity and residual morphology (or features) left after occurrence. Table 2.1 shows the indicative slope angles and peat thicknesses associated with each type.
With time, the features associated with these types of landslide will soften through erosion and drying and re-vegetate, leaving only subtle scars in the landscape (Feldmeyer-Christe and Küchler, 2002; Mills, 2002). A study into vegetation recovery for several UK peat landslide sites indicated that typical features were clearly visible in the field and on aerial photographs for some 30-40 years post failure until full vegetation cover had been achieved. Thereafter, failure morphology degraded and vegetation growth continued until the scars became difficult to identify once around 100 years had elapsed since failure (Mills, 2002).
Table 2.1 Peat landslide types and key controlling parameters (after Dykes and Warburton, 2007a)
|Peat landslide type||Definition||Typical slope range||Typical peat thickness|
|Bog burst||Failure of a raised bog ( i.e. bog peat) involving the break-out and evacuation of (semi-) liquid basal peat||2 - 5˚||2 - 5m|
|Bog flow||Failure of a blanket bog involving the break-out and evacuation of semi-liquid highly humified basal peat from a clearly defined source area||2 - 5˚||2 - 5m|
|Bog slide||Failure of a blanket bog involving sliding of intact peat on a shearing surface within the basal peat||5 - 8˚||1 - 3m|
|Peat slide||Failure of a blanket bog involving sliding of intact peat on a shearing surface at the interface between the peat and the mineral substrate material or immediately adjacent to the underlying substrate||5 - 8˚ (inferred)||1 - 3m (inferred)|
|Peaty debris slide||Shallow translational failure of a hillslope with a mantle of blanket peat in which failure occurs by shearing wholly within the mineral substrate and at a depth below the interface with the base of the peat such that the peat is only a secondary influence on the failure||4.5 - 32˚||< 1.5m|
|Peat flow||Failure of any other type of peat deposit (fen, transitional mire, basin bog) by any mechanism, including flow failure in any type of peat caused by head-loading||Any of the above||Any of the above|
2.2 Controls of peat instability
A number of preparatory factors operate in peatlands which act to make peat slopes increasingly susceptible to failure without necessarily initiating a landslide. Triggering factors change the state of the slope from marginally stable to unstable and can be considered as the 'cause' of failure ( DoE, 1996). There are also inherent characteristics (or preconditions) of some peat covered slopes which predispose them to failure. These are considered further below.
2.2.1 Preparatory factors
The following are some of the transient factors which operate to reduce the stability of peat slopes in the short to medium term (tens to hundreds of years):
(i) Increase in mass of the peat slope through progressive vertical accumulation (peat formation);
(ii) Increase in mass of the peat slope through increases in water content;
(iii) Increase in mass of the peat slope through growth of trees planted within the peat deposit (afforestation);
(iv) Reduction in shear strength of peat or substrate from changes in physical structure caused by progressive creep and vertical fracturing (tension cracking or desiccation cracking), chemical or physical weathering or clay dispersal in the substrate;
(v) Loss of surface vegetation and associated tensile strength ( e.g. by burning or pollution induced vegetation change);
(vi) Increase in buoyancy of the peat slope through formation of sub-surface pools or water-filled pipe networks or wetting up of desiccated areas; and
(vii) Afforestation of peat areas, reducing water held in the peat body, and increasing potential for formation of desiccation cracks which are exploited by rainfall on forest harvesting.
The impacts of factors (i) and (ii) are poorly understood, but the formation of tension cracks, desiccation cracks and pipe networks have been noted in association with many recorded failures. Long-term reductions in slope stability contribute to slope failure when triggering factors operate on susceptible slopes, as described below.
2.2.2 Triggering factors
Peat landslides may be triggered by natural events and human activities. The following natural triggers have been reported in relation to peat instability:
(i) Intense rainfall causing development of transient high pore-water pressures along pre-existing or potential rupture surfaces ( e.g. at the discontinuity between peat and substrate);
(ii) Snow melt causing development of high pore-water pressures, as above;
(iii) Rapid ground accelerations (earthquakes) causing a decrease in shear strength;
(iv) Unloading of the peat mass by fluvial incision of a peat slope at its toe, reducing support to the upslope material; and
(v) Loading of the peat mass by landslide debris causing an increase in shear stress.
Factors (i) and (ii) are the most frequently reported triggers for peat mass movements in the UK. The increasing incidence of multiple peat landslide events may be associated with increased storm frequency (Evans and Warburton, 2007), a climatic trigger considered to be more likely under climate change scenarios.
Triggers associated with human activities include:
(vi) Alteration to natural drainage patterns focussing drainage and generating high pore-water pressures along pre-existing or potential rupture surfaces ( e.g. at the discontinuity between peat and substrate);
(vii) Rapid ground accelerations (blasting or mechanical vibrations) causing an increase in shear stresses;
(viii) Unloading of the peat mass by cutting of peat at the toe of a slope reducing support to the upslope material ( e.g. during track construction);
(ix) Loading of the peat mass by heavy plant, structures or overburden causing an increase in shear stress; and
(x) Digging and tipping, which may be associated with building, engineering, farming or mining (including subsidence).
Natural factors are difficult to control, and while some human factors can be mitigated, some cannot. For these reasons it is essential to identify and select locations and routes for development infrastructure that avoid the deepest peat areas and minimise the impact of the development on peatlands.
Note that Lindsay and Bragg (2004) provide a detailed review of the potential destabilising effects of forestry activities on a peatland in Ireland in association with the Derrybrien failure, including discussion of some of the anthropogenic triggers listed above. In preparing assessments of peat stability, developers should give afforested peatlands (which are often hydrologically disrupted and physically degraded) the same scrutiny as peatlands without forest, even if this may be more arduous in practice (due to concealment of the ground surface by tree cover and associated access difficulties).
The following static or inherited factors may act as preconditions to slope instability in peatlands (Evans and Warburton, 2007; Dykes and Warburton, 2007a):
(i) Impeded drainage caused by a peat layer overlying an impervious clay or mineral base (hydrological discontinuity, especially an iron pan at the base of the peat deposit);
(ii) A convex slope or a slope with a break of slope at its head (concentration of subsurface flow);
(iii) Proximity to local drainage, either from flushes, pipes or streams (supply of water); and
(iv) Connectivity between surface drainage and the peat/impervious interface (mechanism for generation of excess pore pressures).
Dykes and Warburton (2007b) note that "…areas of peat subjected to tine cutting, peat upslope of transverse ditches and thin upland peat on convex mountain slopes should be identified as potentially unstable where not obviously disrupted by previous failures or surface erosion".
2.2.4 Pre-failure indicators
The presence of preparatory or preconditioning factors prior to failure are often indicated by ground conditions that can be mapped or measured remotely or by a site visit. In many cases, sites that have experienced landslides apparently without warning could often have been identified as susceptible to failure by a suitably trained person (see section 1.6) or through relatively inexpensive monitoring strategies. The nature and signs of instability often differ depending on the type and scale of failure. The following critical features are indicative of potential failure in peat environments:
- Presence of historical and recent failure scars and debris;
- Presence of features indicative of tension;
- Presence of features indicative of compression;
- Evidence of 'peat creep';
- Presence of subsurface drainage networks or water bodies;
- Presence of seeps and springs;
- Presence of artificial drains or cuts down to substrate;
- Concentration of surface drainage networks;
- Presence of soft clay with organic staining at the peat and (weathered) bedrock interface; and
- Presence of an iron pan within a mineral substrate.
Each of these indicators is considered below with illustrative plates to guide recognition during site visits.
22.214.171.124 Historical and recent failure scars and debris
The presence of existing landslide scars in a development area may indicate local site conditions conducive to future peat landslide activity. Plates 2.1 and 2.2 illustrate typical peatland morphology associated with historical failure sites. With increasing time since failure, exposed scars will re-vegetate. However, where a bare substrate has been revealed by sliding, full vegetation cover may take 30-40 years to develop.
Although reactivation of the debris or peat surrounding landslide scars has rarely been noted in the published literature, occurrence of peat landslides at different locations on the same hill, separated in occurrence by many years, has been identified on several occasions. Therefore, the existence of a peat landslide scar in a development area provides a strong indicator of potential future peat landslide hazard (see Dykes and Warburton (2007a) for oblique aerial photographs of the peat landslide types shown in Table 2.1).
126.96.36.199 Tension features
Cracking of the upper layers or full thickness of peat may indicate an accumulation of stress in peat soils as well as provide a route for surface water to infiltrate rapidly to basal peat layers and contribute to the generation of excess pore-water pressures. Tension features may include tension cracks, which are narrow and deep fissures, frequently infilled with water, and which may be continuous or discontinuous for several tens of metres ( Plate 2.3). Alternatively, shallow tears, which are wider and shallower 'diamond' shapes may indicate tension at the surface only ( Plate 2.4). Concentric tiers of arcuate tension cracks may indicate local displacement, while multiple intersecting cracks may be a precursor to fragmentation of the peat into rafts or blocks ( Plate 2.5).
188.8.131.52 Compression features
Compression features usually indicate displacement upslope which has resulted in the formation of ridges ( Plate 2.6), thrusts ( Plate 2.7) or extrusion features ( Plate 2.8). Often, related tension features will be visible at the upslope limit of peat displacement.
184.108.40.206 Peat creep
Tension and compression features can be associated with creep of the peat blanket on a slope. Zones of tension are often juxtaposed with compression ridges in response to creep of the peat mass and changes in local slope gradient. At the surface such movements can be detected by displacement of walls and boundaries and tilting of fences and posts.
220.127.116.11 Subsurface drainage networks or water bodies
Subsurface drainage pathways may enable high transient pore-water pressures to develop under conditions of enhanced water supply, e.g. during an intense rainstorm. Soil piping is widespread in upland blanket peat catchments in the UK, with pipes tending to be more prevalent at hillslope summits and footslopes and in areas of peatland subject to moorland gripping (Holden, 2004; 2005).
Such pipe drainage networks ( Plate 2.9) can often only be identified on-site by the sound of running water beneath the surface. Rarely, a pipe ceiling may collapse, leaving a hole in the peat surface ( Plate 2.10). If pipe networks are identified, their size and extent can be ascertained through non-intrusive ground investigation e.g. Ground Penetrating Radar (Holden, 2004; see section 18.104.22.168).
Larger subsurface water bodies, formed where pipes have become blocked or where spring lines are present below the surface, can be identified by 'trembling' at the peat surface. Continued supply of water (without release) to a subsurface water body may cause visible swelling of the peat mass over periods ranging from a few hours to several months. Evidence of drainage outlets should also be noted as these usually indicate a well-developed subsurface pipe network. The presence of sediment discharged at natural pipe outlets often indicates a deep subsurface drainage network with periodically high water pressures.
22.214.171.124 Seeps and springs
Groundwater seeps and springs are controlled by seasonal rainfall. Large fluctuations in rainfall may increase the rate of groundwater discharges and if this occurs following a period of drought there may be an increased likelihood of peat landslide occurrence.
126.96.36.199 Artificial drainage and cuttings
Moor drains in open peatlands (or grips), forest drains on afforested peatlands and cuttings (for fuel) to substrate, particularly where oriented transverse or oblique to the slope contour, may weaken a peat covered slope. The drain or cut will create a vertical discontinuity, removing tensile strength in the upper layers and enabling ponding of water.
188.8.131.52 Drying and cracking
Drying of peat caused by periods of drought or by drainage (natural or man-made) may also cause cracking, providing pathways for rapid infiltration of water to horizons at depth within the peat mass. Crack networks can also be caused by drying of a peat mass associated with forestry and the associated drawdown of the water table. These crack networks can be found both under and between lines of trees in a forest stand, although they may be hidden under a litter layer (see Lindsay and Bragg, 2004).
184.108.40.206 Concentration of natural drainage networks
Natural surface drainage pathways also provide a means of supplying water to a susceptible peat area. These may be manifest as gullies ( Plate 2.11), which are typically steep sided and may have bare peat floors or be fully vegetated around a single thread sinuous stream. Alternatively, flushes ( Plate 2.12) may enable significant diffuse flow of water without an obvious single thread watercourse.
Areas of peat which are already dissected by gully networks (with numerous haggs and groughs) have rarely been associated with peat instability, however, flush zones are often cited as coincident with the locations of peat instability.
220.127.116.11 Soft clays
Landslides in which peat is the major part (by volume) may fail within the peat deposit, at the interface between it and the peat substrate, or within the substrate underlying the peat (Dykes and Warburton, 2007a). Glacial tills with a high clay content have been cited as the substrate for a number of failures where the shear surface was interpreted to be within the substrate. The uppermost clay layers beneath peat may be weakened by organic acids leached from the base of the peat, producing particularly soft clays. Where clays have been softened in this way, augering or probing may penetrate the clays with a similar refusal to the overlying peat, and if this occurs, it should be noted.
18.104.22.168 Iron pans
A number of studies have identified impermeable iron pans (or similar hardened horizons formed by leaching) in the upper centimetres of the mineral substrate beneath peat deposits. These pans may have smooth, almost 'polished' surfaces and provide little frictional resistance to the overlying peat.
Any of the indicators described in 22.214.171.124 to 126.96.36.199 may in isolation indicate future potential for peat landslides to occur and combinations of these features may indicate a greater susceptibility to failure. It should be noted that there are few citations in the academic literature that support the idea of increasing slope gradient and increasing peat thickness (particularly together) as the primary controls on failure. Both Evans and Warburton (2007) and Boylan et al. (2008) note that the majority of recorded failures are on relatively low gradients (typically 4-8°) where peat accumulation may be greater and in thin to moderate thickness peats (typically 0.5 - 2.0m deep in blanket peat, but thicker in raised bogs; Boylan et al., 2008).
Email: Energy Consents Unit
Phone: 0300 244 4000 – Central Enquiry Unit
The Scottish Government
St Andrew's House