Chapter 1 Introduction

R.G. Cooper

Mass movements in context

Mass movements in the British context

Jones and Lee (1994) describe 'mass movement' as 'a broad spectrum of gravity[-driven] slope movements', of which the larger discrete movements are generally described as 'landslides'.

These mass-movement phenomena are a major influence on much of the landscape of Great Britain, but vary considerably in scale. Some mass-movement processes are shallow (operating near the land surface), slow, and affect large areas. For example, 'soil creep' has been taking place on nearly all terrestrial slopes since the retreat of glaciers during the Devensian Stage (the last glacial period of the Pleistocene Epoch in Britain, which ended about 11 500 calendar years ago). Similarly, many landscapes (e.g. Dartmoor) are mantled by 'solifluction' sheets of sediment (slow downslope-moving saturated soil or rock debris), the process that created them usually being ascribed to former periglacial (tundra-like) climatic conditions.

At the other end of the scale are deep-seated 'landslides', whose occurrence under present climatic conditions in Britain is relatively rare both areally and temporally (except on particular stretches of the east and south coasts). Many of these have clearly taken place in the past under conditions more conducive to mass movements and are very widespread. These mass-movement features are the principal subject of the present volume.

A study undertaken in 1984–1987 for the former Department of the Environment (DoE) by Geomorphological Services Ltd (GSL; published in 1988) in association with Rendel Palmer & Tritton, that produced an inventory of 8835 landslides in Great Britain has been analysed by Jones and Lee (1994). A major conclusion drawn from the analysis is that most inland landslides in Great Britain are relict but dormant (i.e. capable of being re-activated by engineering works, building or other disruptive activities). In contrast, coastal landsliding is a present-day process, possibly associated with rising sea level and drainage.

A particular value of the inventory has been the provision of information for local and regional planners (Clark et al., 1996), who have to deal with the consequences of landslides — present-day, recent and relict — in relation to land-development applications.

No other Geological Conservation Review (GCR) volume has had the benefit of such a major survey of the features with which it is concerned carried out by another organization at a critically important time. The survey took place around the time of the period of GCR fieldwork in the 1980s. The GSL survey, however, was concerned with landslides that have been mentioned or shown in documents. Therefore it does not purport to be a complete inventory of known landslides in Great Britain. The distribution of landslides identified by the survey, described as 'ancient' and 'youthful', is shown in (Figure 1.1).

The present writer (RGC) was involved as a collector of data for north-east England in the GSL exercise, which led to the production of the distribution map (Figure 1.1). It was clear that when plotted on a 1:125 000 scale map, the distribution of reported landslides in north-east England alone was likely to be a very poor representation of the true distribution of landslides actually identifiable in the field. A major reason for this was that landslides were not recorded equally well in the different surveyed areas, creating apparent, but not actual, demarcations of high- and low-density areas of landslides (see below). The demarcations, as recorded in the literature, often correlated to the boundaries between the various map areas of individual British Geological Survey maps, memoirs, and Mineral Assessment Reports. Examples included a cluster of landslides in North Yorkshire immediately west of Ripon, another around Barnard Castle in County Durham, and a group around Bellingham in Northumberland. Since it is unlikely that landslide density correlates to British Geological Survey map-sheet areas, this must indicate unevenness in the documentation between the sheets, memoirs or reports for adjacent areas. There are several reasons for such unevenness:

1. It is clear that for areas surveyed up to some time in the 1930s, landslides were simply not marked on the resulting published British Geological Survey sheets. Examination of the one-inch sheets of the North York Moors area produced from surveys made by C. Fox-Strangways between 1880 and 1910 reveals no landslides at all. Yet the six-inch maps from which they were compiled show a very large number of slides. These do not have marked boundaries: the word 'slip' is written across the relevant slope; nevertheless, they were identified, marked and recorded by the surveyor.
2. Even though it has been British Geological Survey practice to mark landslides on the one-inch and 1:50000 sheets, some sheets specifically exclude them. These include one-inch sheet 50 (Hawes), published in 1971, which has a note appended to the legend: 'N.B. Landslips are not indicated on this map'.
3. The interests and aptitudes of each surveyor have had an influence. About half of the currently available British Geological Survey memoirs have 'landslip' in their indexes. Where they do not, either there are none in the area concerned, or the surveyors were not interested in such phenomena. The latter is the present writer's (RGC's) explanation of the low number of recorded landslides in Nottinghamshire (seven in (Table 2.2) of Jones and Lee, 1994). The county was surveyed by geologists selected for their expertise in coal and economic geology; landslides were, therefore, not a major concern to them. The memoirs that they produced concentrate heavily on phenomena at depth; surficial geology is assigned a very minor role.

Taking a national view, even where 'landslip' can be found in the index of a British Geological Survey memoir, recorded coverage of such mass movements is very variable. The area under consideration may contain a single large and obvious slide, recorded as such and which a surveyor could hardly fail to mention, but smaller movements in the region may be overlooked or neglected. Alternatively, an area may have no major mass movements, but the surveyor may have a particular interest in landslides, perhaps because of Quaternary research interests, and so may record comparatively many more occurrences. Differences in the interests and aptitudes of surveyors is the only tenable explanation of why some of the recently surveyed one-inch and 1:50 000 sheets of Northumberland are replete with landslides (sheet 13 (Bellingham) has 71) while some of the adjacent sheets, with similar geology and comparable terrain — but different surveyors — show none at all.

Similar observations were made about other parts of the country by other collectors of data for the exercise, leading Jones and Lee (1994) to observe that 'the patchiness of the distribution raises questions as to the extent to which the concentrations displayed in the map [here (Figure 1.1)] reflect the true pattern of landslides on the ground as against spatially variable reporting'. They continue:

'It now seems certain that the pattern merely highlights those landslides which happen to have been investigated, mapped and reported, and the extent to which the total available corporate knowledge of landsliding was tapped by the survey. It is undoubtedly true that many reports of landslides published in obscure journals and old newspapers were not accessed by the survey, and the same is true of the data held in the files of numerous individual professionals, companies and even some national organizations. It must also be stressed that there must be numerous other landslides that have not yet been recorded because they exist in remote areas, are concealed by woodland, are relatively insignificant or have yet to be actually recognized as landslides. This is clearly illustrated by the results of the … Applied Earth Science Mapping of the Torbay area (1988) which raised the total of known and reported landslides from 4 to 304. Even in the South Wales Coalfield, which has been the subject of a major landslide inventory exercise by the British Geological Survey, a detailed mapping programme in the Rhondda valleys resulted in an increase in the number of recorded landslides from 102 to 346. Clearly, in some areas, the harder you look the more examples you find. Indeed, extrapolation leads to the inevitable conclusion that the actual number of landslides in Great Britain is many times in excess of the 8835 recorded so far by this survey.'

The Torbay study referred to is described in Geomorphological Services Ltd (1988) and Doornkamp (1988). As stated by Jones and Lee (1994), the pattern of landslides displayed on the map (Figure 1.1) must be treated with caution in that it reflects under-representation of the true pattern, as an artefact of investigative interests and recording bias.

Mass movements in the European context

A large amount of research has been carried out on mass movements in Europe, particularly in relation to three broad factors: climate, topography and geology. It is worth noting, however, that this tripartite division does not create exclusive, distinct categories. Geology and topography, in particular, are intimately linked, with the geology (lithology and structure) controlling the topography in some detail. Also, few mass movements can be ascribed to a single causal factor, or even to a single type of causal factor.

(a) Climatic factors

Two initiatives by the European Commission (EC) have been concerned with Europe-wide collection and analysis of data on mass movements: the EPOCH project (Temporal Occurrence and Forecasting of Landslides in the European Community) and the TESLEC project (The Temporal Stability and Activity of Landslides in Europe with Respect to Climatic Change). The EPOCH project collected data on the past occurrence and frequency of landsliding in Europe. The UK EPOCH team then went on to extract from these data changes of geomor-phological activity that may be related to climatic change in the last 20 000 years. These results are summarized in (Figure 1.2) for Holocene times based on dates of named landslides in the UK and for selected European countries, respectively (Brunsden and Ibsen, 1997; Ibsen and Brunsden, 1997). They suggest that landslide activity may be related to specific climatic periods and that the existing knowledge of this could be substantially improved.

The TESLEC project has been concerned with the effects and modelling of climate change on mass movements in Europe. This has involved continued collation of data on the past distribution of landslides. This work has shown that there are few decades when landslide events have not occurred in certain regions such as the Spanish Pyrenees and Barcelonette in the French Alps (Brunsden et al., 1996a).

The 'Landslide Recognition' survey (Dikau et al., 1996; the production of the survey was an initial objective of the TESLEC project), and the DoE review revealed that in Great Britain the major problem with respect to climate change scenarios is the potential for the re-activation of dormant landslide complexes, rather than the potential for first-time slides. By far the biggest potential problem is the possibility that climate change might generate widespread movement in the very large landslide complexes that lie at the foot of many of the escarpments in Mesozoic strata. These complexes are, however, rather rare on Chalk, for example the Chiltem Hills, Salisbury Plain and the Marlborough Downs, and along the North and South Downs. However, where clay is exposed at the base of a slope in Chalk strata, occasional landslide complexes are to be found, for example the Castle Hill landslide at the entrance of the Channel Tunnel, the coastal termination of the North Downs at Folkestone Warren (see Chapter 7), the Dorset coast, and inland at Birdsall Brow in North Yorkshire.

(b) Topographical factors

The investigations of climatic factors also required other influences on mass movements to be identified, such as unloading, sea-level rise, seasonal ground freezing, and caprock loading (Brunsden and Ibsen, 1997). However, the European chapters of a worldwide survey of the extent and economic significance of landslides (Brabb and Harrod, 1989) point to the over-riding importance of individual high-intensity weather events, rather than climatic trends, as a precipitating factor for many mass movements. As stated, this has happened in Great Britain, but only rarely. However, a landslip in Yorkshire in 1755 took place 149 days after a very high-intensity, short-duration, small area rainfall event. The reason for this time-lag is not known, but it is within the range of flow-through times from precipitation of water on the ground surface to its emergence at groundwater-fed springs in the area. An earthquake can probably be ruled out as the immediate trigger of this landslide (Cooper, 1997).

France (Flageollet, 1989) has important topographical differences from Great Britain, including relatively new mountain ranges with steep, high slopes, such as the Alps and the Pyrenees. This topography leads to a similar range of types of mass movements to Great Britain, but the proportions are different. Particularly instructive are the variations of style of failure along the Bessin Cliffs on the north coast of Normandy, at Pointe du Hoc, Raz de la Percee, le Bouffay, le Chaos and Cap Manvieux (Maquaire and Gigot, 1988).

(c) Geological factors

Generally, where areas of Europe have geological situations not found in Great Britain, they also have types of mass movement not found there. The most obvious case is the quickclay deposits of Norway and Sweden, where marine clays deposited during Holocene times have been uplifted isostatically to become part of the land surface (Gregersen and Sandersen, 1989). Since emerging from the sea, these deposits have been subject to subaerial erosion, and the salt water has leached out from the soil matrix. This leaching increases the sensitivity (s_t) of the clay from, typically, an s_t value of 3–6, to a value greater than s_t = 20. When the salt content falls below 1 gl^–1, the clay becomes a quickclay. An unleached marine clay remains plastic on re-moulding, but a quickclay can transform into a liquid (Bjerrum, 1954; Bjerrum et al., 1969).

Norway also has a large area of hard-rock mountains, liable to rockfalls and rockslides, rather like the Scottish Highlands.

Mass movements in the global context

From the global perspective, mass movements in Great Britain are unremarkable for their size, frequency, the hazards they pose, and their overall variety. They are, perhaps, remarkable for the small proportion that are currently active, and conversely for the large proportion that are generally attributed to past climatic conditions rather than those of the present day, in particular periglacial conditions and immediate post-glacial conditions.

The reasons for this limited manifestation of mass movements here are not hard to find. The British Isles are not located close to a tectonic plate boundary, and have not been subject to volcanic activity or significant seismic activity for many millions of years, so these potential triggering mechanisms do not play a major role. Isostatic re-adjustment to the melting of the most recent Quaternary glaciers, and retreat of the last ice-sheet, may still produce minor earthquakes, but these are infrequent, and hardly of significance even in the triggering of the few active mass movements that are located in the Scottish Highlands.

The principal limiting factor for British mass movements seems to be available relief. The highest point in Great Britain, the summit of Ben Nevis, reaches only 1343 m above OD; long, steep slopes are, therefore, something of a rarity. However, Great Britain's steep slopes in upland areas have been the sites of debris flows, but comparatively only a few rockfall avalanches.

Great Britain's temperate maritime climate has seldom produced extremes of rainfall capable of giving rise to a large-scale spate of mass-movement events in an area over a timescale of a few hours. This has been known to happen, for example on Exmoor in August 1952 (Gifford, 1953; Delderfield, 1976), but such events are very uncommon, although climate change may increase their frequency in the future. Likewise, although solifluction has been an important mass-movement process across much of Great Britain, it is only in the extremes of upland Britain that conditions are sufficiently cold for periglacial processes such as solifluction to be currently taking place (Ballantyne and Harris, 1994).

On the other hand, Great Britain has an immense variety of landforms, which occur on bedrock of varied geological age and reflect differences in lithology (rock-type) and geological structure (such as faults and folds). This variety has given rise to slopes that, in valley-sides, expose rocks of greatly varying resistance to erosion, and so produce a variety of degrees of slope steepness. Many slopes of the upland areas have been steepened relatively recently by glacial erosion. Likewise many kilometres of the British coast consist of vertical or sub-vertical cliffs, of variable height. As would be expected, these features have led to the development of a great number of mass-movement sites. While the majority are unremarkable, collectively they demonstrate a variety of features associated with Quaternary erosion, scarp retreat, and landscape shaping.

There is, however, one aspect of the mass movements in Great Britain that has had a substantial, and possibly disproportionate, influence globally. This is their role in the development of knowledge of mass movements, their mechanisms and their countermeasures (Hutchinson, 1984). Thus, one of the mass-movement sites chosen for the GCR includes, arguably, the first ever large-scale landslide to be described by geologists, the Bindon landslide, part of the Axmouth–Lyme Regis mass-movement GCR site. In addition, a long series of studies of the behaviour of London Clay (Eocene-age deposits) has illuminated the mass-movement behaviour of all clay strata. More recently, the recognition of toppling as a separate and distinct type of slope failure has depended upon the study of British examples.

Classification of mass-movement types

The classification of mass movements into types has attracted much attention since the suggestions made in 1938 by Sharpe. Classification is dealt with in some detail in the present chapter. However, characterizing landslide type, while important scientifically, has not been the sole consideration in the selection of mass-movement GCR sites, some of which were selected on the basis of the presence at a site of an atypical or otherwise particularly interesting feature or group of features.

The classification system of mass-movement features adopted for the purposes of selecting mass-movement GCR sites in the 1980s, was originally that of Hutchinson (1968a), the overall breadth of which, including creep, frozen-ground phenomena and landsliding, indicated a convenient scope to adopt for the term 'mass movement' (Table 1.1)a. Hutchinson (1968a) makes a most significant point about mass movement: 'mass movements exhibit great variety, being affected by geology, climate and topography, and their rigorous classification is hardly possible'.

Despite this general proviso, several classifications were published in the 1970s, 1980s and 1990s, including those of Zaruba and Mencl (1969), Varnes (1978), Brunsden (1979), Selby (1982), Geomorphological Services Ltd (in Jones and Lee, 1994), Hutchinson (1988; see (Table 1.1)b) and most notably The Multilingual Landslide Glossary developed by the International Geotechnical Societies' UNESCO Working Party for World Landslide Inventory (WP/WLI, 1993; see below, where the glossary is reproduced in full). Where possible the recommendations and terminology of this last-mentioned group are now used throughout this volume in order to follow international practice.

All such classifications are to some extent imperfect, in that any classification of mass movements is essentially trying to divide a continuum into classes, raising the obvious difficulty of locating the distinguishing boundaries between types. Furthermore, many sites incorporate a number of different features of a variety of mass-movement types belonging to different classes. Thus placing sites into a particular category is subject to opinion.

The introduction to Chapter 2 of the present volume addresses this difficulty in respect of the old hard rocks of the British mountains, adapting the Hutchinson (1988) schema to more specific circumstances.

Arguably, the classification used by civil engineers gives perhaps the most clearly defined and separate 'types', as it is a classification not of mass-movement types, but of failure types (e.g. in Hoek and Bray, 1977).

The Multilingual Landslide Glossary

The Multilingual Landslide Glossary is an international standard for the description of landslides (WP/WLI, 1993; Cruden et al., 1994). Its English version is given here in full (see also Dikau et al., 1996). The glossary is available in Arabic, Chinese, English, French, German, Hindi, Italian, Japanese, Persian, Russian, Spanish, Sinhala, and Tamil. While giving a comprehensive glossary of terms for the various features of a landslide (Figure 1.3) and (Figure 1.4), it also divides landslides into five types: fall, topple, slide, spread and flow (Figure 1.5). Each of these types is modified by other qualities, of which two are of particular relevance to the classification: distribution of activity (seven qualifiers of type; (Table 1.1)b, (Figure 1.6)), and style of activity (five qualifiers of type; (Figure 1.7)). This leaves two that are of less relevance to classification of 'type': dimensions (Figure 1.8), and state of activity (see (Figure 1.5)).

Therefore there are 175 (5 x 7 x 5) theoretically possible types. Of these, the editors and contributors to Landslide Recognition (Dikau et al., 1996) choose to describe 15, which leaves one to speculate on the actual existence of field examples of the remaining 160.

(Table 1.1) a. Hutchinson's classification of mass movements on slopes b. Hutchinson's (1988) classification (first two levels only)

a.

b.

Landslide features (Figure 1.3)

1. Crown: the practically undisplaced material still in place and adjacent to the highest parts of the main scarp (2).
2. Main scarp: a steep surface on the undisturbed ground at the upper edge of the landslide, caused by movement of the displaced material (13) away from the undisturbed ground. It is the visible part of the surface of rupture (10).
3. Top: the highest point of contact between the displaced material (13) and the main scarp (2).
4. Head: the upper parts of the landslide along the contact between the displaced material (13) and the main scarp (2).
5. Minor scarp: a steep surface on the displaced material (13) of the landslide produced by differential movements within the displaced material (13).
6. Main body: the part of the displaced material (13) of the landslide that overlies the surface of rupture (10) between the main scarp (2) and the toe of the surface of rupture (11).
7. Foot: the portion of the landslide that has moved beyond the toe of the surface of rupture (11) and overlies the original ground surface (20).
8. Tip: the point on the toe (9) farthest from the top (3) of the landslide.
9. Toe: the lower, usually curved margin of the displaced material (13) of a landslide; it is the most distant margin of the landslide from the main scarp (2).
10. Surface of rupture: the surface that forms (or which has formed) the lower boundary of the displaced material (13) below the original ground surface (20).
11. Toe of the surface of rupture: the intersection (usually buried) between the lower part of the surface of rupture (10) and the original ground surface (20).
12. Surface of separation: the part of the original ground surface (20) overlain by the foot (7) of the landslide.
13. Displaced material: material displaced from its original position on the slope by movement in the landslide. It forms the depleted mass (17) and the accumulation (18).
14. Zone of depletion: the area of the landslide within which the displaced material (13) lies below the original ground surface (20).
15. Zone of accumulation: the area of the landslide within which the displaced material (13) lies above the original ground surface (20).
16. Depletion: the volume bounded by the main scarp (2), the depleted mass (17), and the original ground surface (20).
17. Depleted mass: the volume of the displaced material (13) which overlies the surface of rupture (10) but underlies the original ground surface (20).
18. Accumulation: the volume of the displaced material (13) which lies above the original ground surface (20).
19. Flank: the undisplaced material adjacent to the sides of the surface of rupture (10). Compass directions are preferable in describing the flanks, but if left and right are used, they refer to the flanks as viewed from the crown (1).
20. Original ground surface: the surface of the slope that existed before the landslide took place.

Landslide dimensions (Figure 1.4)

1. The width of the displaced mass, Wd, is the maximum breadth of the displaced mass perpendicular to the length of the displaced mass, Ld (4).
2. The width of the rupture surface, Wr, is the maximum width between the flanks of the landslide, perpendicular to the length of the rupture surface, Lr (5).
3. The total length, L, is the minimum from the tip of the landslide to the crown.
4. The length of the displaced mass, 14, is the minimum distance from the tip to the top.
5. The length of the rupture surface, Lr, is the minimum distance from the toe of the surface of rupture to the crown.
6. The depth of the displaced mass, Dd, is the maximum depth of the displaced mass, measured perpendicular to the plane containing Wd (1) and Ld (4).
7. The depth of the rupture surface, Dr, is the maximum depth of the rupture surface below the original ground surface measured perpendicular to the plane containing Wr (2) and Lr (5).

Types of landslides (Figure 1.5)

1. A fall starts with detachment of soil or rock from a steep slope along a surface on which little or no shear displacement takes place. The material then descends largely through the air by falling, saltation or rolling.
2. A topple is the forward rotation, out of the slope, of a mass of soil or rock about a point or axis below the centre of gravity of the displaced mass.
3. A slide is the downslope movement of a soil or rock mass occurring dominantly on surfaces of rupture or relatively thin zones of intense shear strain.
4. A spread is an extension of a cohesive soil or rock mass combined with a general subsidence of the fractured mass of cohesive material into softer underlying material. The rupture surface is not a surface of intense shear. Spreads may result from the liquefaction or flow (and extrusion) of the softer material.
5. A flow is a spatially continuous movement in which surfaces of shear are short-lived, closely spaced and usually preserved. The distribution of velocities in the displacing mass resembles that in a viscous fluid.

States of activity of landslides (Figure 1.6)

1. An active landslide is currently moving; the example in (Figure 1.6) shows that erosion at the toe of the slope causes a block to topple.
2. A suspended landslide has moved within the last 12 months, but is not active (1) at present; the example in (Figure 1.6) shows local cracking in the crown of the topple.
3. A re-activated landslide is an active (1) landslide which has been inactive (4); the example in (Figure 1.6) shows that another block topples, disturbing the previously displaced material.
4. An inactive landslide has not moved within the last 12 months and can be divided into four states: (5) dormant, (6) abandoned, (7) stabilized, and (8) relict.
5. A dormant landslide is an inactive (4) landslide which can be re-activated (3) by its original causes or by other causes; the example in (Figure 1.6) shows that the displaced mass begins to regain its tree cover, and scarps are modified by weathering.
6. An abandoned landslide is an inactive (4) landslide which is no longer affected by its original causes; the example in (Figure 1.6) shows that fluvial deposition has protected the toe of the slope; the scarp begins to regain its tree cover.
7. A stabilized landslide is an inactive (4) landslide which has been protected from its original causes by remedial measures; the example in (Figure 1.6) shows that a wall protects the toe of the slope.
8. A relict landslide is an inactive (4) landslide which developed under climatic or geomorphological conditions considerably different from those at present; the example in (Figure 1.6) shows that uniform tree cover has been established.

Distribution of activity in landslides (Figure 1.7)

Section 2 in each part of (Figure 1.7) shows the slope after movement on the rupture surface indicated by the shear arrow in the

1. In an advancing landslide the rupture surface is extending in the direction of movement.
2. In a retrogressive landslide the rupture surface is extending in the direction opposite to the movement of the displaced material.
3. In an enlarging landslide the rupture surface of the landslide is extending in two or more directions.
4. In a diminishing landslide the volume of the displaced material is decreasing.
5. In a confined landslide there is a scarp but no rupture surface visible at the foot of the displaced mass.
6. In a moving landslide the displaced material continues to move without any visible change in the rupture surface and the volume of the displaced material.
7. In a widening landslide the rupture surface is extending into one or both flanks of the landslide.

Styles of landslide activity (Figure 1.8)

1. A complex landslide exhibits at least two types of movement (falling, toppling, sliding, spreading and flowing) in sequence; the example in (Figure 1.8) shows gneiss and a pegmatite vein toppled with valley incision. Alluvial deposits fill the valley bottom. After weathering had weakened the toppled material, some of the displaced mass slid farther downslope.
2. A composite landslide exhibits at least two types of movement simultaneously in different parts of the displacing mass; the example in (Figure 1.8) shows that limestones have slid on the underlying shales causing toppling below the toe of the slide rupture surface.
3. A successive landslide is the same type as a nearby, earlier landslide, but does not share displaced material or rupture surface with it; the example in (Figure 1.8) shows that the latter slide, AB, is the same type as CD, but does not share displaced material or a rupture surface with it.
4. A single landslide is a single movement of displaced material.
5. A multiple landslide shows repeated development of the same type of movement.

GCR site selection

Methodology

The rationale, methodology and history of the selection of sites for inclusion within the Geological Conservation Review programme has been discussed in detail by Wimbledon et al. (1995) and in the introductory GCR volume (Ellis et al., 1996). The main factors considered during the selection process can be summarized as:

1. importance to the international Earth scientist community;
2. presence of exceptional (classic, rare or atypical) geological/geomorphological features; and
3. national importance for features that are representative of geological events or processes that are fundamental to understanding the geological/geomorphological history of Great Britain.

There are also the principles in GCR site selection that a chosen site should be the best available example of its kind, and that there should be a minimum of duplication of features between GCR sites.

To adapt these criteria specifically to mass movements has been particularly difficult, compared to geological (rather than geomorphological) selection categories.

Given Hutchinson's (1968a) classification, one particular 'type' of mass movement might be represented by several sites to show the different circumstances in which that 'type' typically occurs. However, during original GCR site selection it was not envisaged that, for example, the 'type' called 'rotational slips' would be represented in the ultimate GCR register by an example in strata from each of the geological periods, or in each region of the country, in which that type is found. Using the GCR ethos (Wimbledon et al., 1995; Ellis et al., 1996), the method of GCR site selection followed for mass movements was that set out below.

1. A first-tranche list of 23 candidate GCR sites was assembled, following literature survey and initial research. This list was circulated to relevant members of the geological, geomorphological and civil engineering communities, with the suggestion that they might delete some sites from the list, and recommend other sites not on the list.
2. The result was that 116 candidate sites were suggested for selection for the GCR by the consultees (Table 1.2) — a five-fold increase in the originally circulated list. A statistical summary was produced, which was published in February 1982 in Earth Science Conservation (Anon. [Cooper] in Black, 1982). The text of part of this article is reproduced here:

...Of these [sites suggested for consideration], 65% are located in England (13% South-East, 22% Midlands, 20% North), 24% are located in Scotland, and 11% in Wales. One third have coastal locations, and 14% are on offshore islands. Just over a quarter of the sites suggested are in Carboniferous rocks, with the Namurian of the central and southern Pennines prominent. As might be expected, the scarp-and-vale topography of the Jurassic is also a major location of recommended features (23%). Other important locations are the Precambrian (14%) and the Cretaceous (12%). Sites in the Devonian and Quaternary each make up 6% of the total, while Cambrian, Triassic and Quaternary sites each provide 3%. Permian, Silurian and Ordovician sites each provide less than 2% of the total.

The responses to the postal survey exposed several general problems associated with site selection. Firstly, there is the problem of the transience of most medium-and small-scale mass movement phenomena. Features which yield valuable information and are educationally instructive immediately after the mass movement has taken place, may after a few years become totally obscured by the smaller-scale processes that tend to even-out irregularities on slopes. In other words, the value of such sites often resides in their freshness. There would be little point in selecting for conservation sites which are unlikely to persist, since, unlike quarry sections, mass movement sites can seldom be 'cleaned up' without destroying those features in which their academic interest resides. Secondly, the well-known mass movement classifications of Hutchinson and of Varnes include types which are either of little importance or absent altogether in Great Britain. Some correspondents expressed the view that coverage should include as wide a range of examples of mass movement types as possible, while others suggested that only sites with a pronounced morphological expression, either on the surface or in section, should be considered. A third problem is that many continuously operating, small-scale processes, which are of great importance in Great Britain, do not give rise to recognizable features either on the surface or in section. For this reason they are not readily conservable. However, such processes, for example soil creep, are so widespread and commonplace that they are deemed not to require conservation at a few, specified, 'representative' sites. A pragmatic solution has been adopted, whereby, as far as possible, Great Britain's 'best' example of each mass movement type is only to be selected if that example is also a 'good' one when viewed from a global perspective.

Of the individual sites recommended in the responses to the postal survey, the complex of rotational slips on the south coast between Axmouth and Lyme Regis was the most often mentioned, closely followed by the slumps in Quaternary deposits on the north Norfolk coast around Cromer and Trimingham. Next most frequently suggested was Warden Point on the coast of the Isle of Sheppey, followed jointly by Folkestone Warren in Kent, the solifluction lobes on the Lower Greensand escarpment near Sevenoaks, the Undercliff on the south coast of the Isle of Wight, the slips at Chale Bay on the Isle of Wight, and the massive features at Quiraing and The Storr, in the Trotternish peninsula of the Isle of Skye. Other much-mentioned sites included High Halstow in Kent, the area around Bath, screes in the Lake District and near Llangollen, Mam Tor in the Peak District, and the abandoned cliff at Hadleigh Castle in Essex. Even between these most mentioned sites, there are obvious overlaps of mechanism, of surface form, and of cross-sectional features, so that the inclusion of all of them in the Review would involve unwarranted duplication. Conversely, several sites have already been assessed as suitable for inclusion, even though each was only suggested by a single correspondent.

At the close of the 1981 field season, 62% of the recommended sites had been either examined in the field, or excluded from the exercise without a visit. The latter course has been taken either through prior personal knowledge of the site in question, or because a reading of the literature shows the site to be an inferior duplicate. Many sites have been examined in aerial view at the Cambridge University Collection of Air Photographs. This has proved a most valuable aid to site selection or elimination. It is anticipated that between twenty and thirty sites will eventually be selected for inclusion in the Review; the final choice awaits completion of the programme of field visits.'

(Table 1.2) The candidate mass-movement GCR sites suggested by the panel of experts consulted in the 1980s.

3. After sifting the original and emended lists, the remaining sites were all visited in the field during 1981 and 1982 by the present author (RGC). Photographs were taken along with some measurements, facilitating direct comparison of what might be termed 'competitive' candidate sites (the GCR being a minimalist scheme, see Ellis et al., 1996). The aim at this stage was to ensure that a complete network of sites was established to represent the variety of mass-movement types and forms found in Great Britain. After consultation and revision, a list of 28 sites was finally produced; this list, with short descriptions, has been published in Jones and Lee (1994, pp. 242–7). This is the list (Table 1.3) that was finally adopted, and is described in the present book, with the exception of the site at Spot Lane Quarry near Maidstone in Kent, described by Worssam (1963), which was included as an example of strata exhibiting two superficial structures: cambering and gulling. However, between 1980 when the site was visited, and a return visit in 1996, a housing estate had been extended onto the area concerned. A small exposure has been preserved there on account of the fossil fauna of a gull filling (selected for the Quaternary of South-East England GCR Block), but otherwise this remnant exposure now shows camber and gull features no better than many other sites across the country.

(Table 1.3) The final list of selected mass-movement sites as drawn up in the early 1980s.

GCR Editor's note:

A review of the Scottish Highland mass movements carried out after Roger Cooper's death showed that there were eight sites, which, as a result of recent investigations by Colin Ballantyne and David Jarman, met GCR standards. The new sites (described in Chapter 2) are listed in (Table 1.4). Had this information been available at the time of the original scoping exercise, when none of these sites were suggested, there is little doubt that they would have been included. The review also showed that applying the 'minimalist' principle, one site, Glen Pean, would not now have been included in the GCR. Revised site information also became available for several of the already selected Scottish sites (Coire Gabhail — Chapter 4; and the Trotternish Escarpment (Quiraing and The Storr) — Chapter 6). (Table 1.5) and (Table 1.6) have been revised to recognize these changes.

(Table 1.4) The supplementary sites added to the GCR following recent research in Scotland.

Site classification

The style and type categories from The Multilingual Landslide Glossary, with the codes from Hutchinson's classifications of 1968a and 1988, are shown along-side brief descriptions for each of the mass-movement sites selected for the GCR, in (Table 1.5). However, the GCR deals with sites (areas of land with a defined boundary), and the classifications deal with the types of movement involved in a displaced mass, or mass undergoing displacement. Thus, Warden Point, for example, is recorded as composite in style, involving both sliding and toppling. This could give the misleading impression that at Warden Point mass-movement events characteristically involve both toppling and sliding together. In fact, Warden Point shows the results of several mass-movement events, side by side along the coast. Of these, most are slides, but one shows toppling.

(Table 1.6) shows the mass-movement GCR sites described in the present volume, classified in two ways. First, by the stratigraphical order of the major geological systems in which the mass-movement phenomena occur in Great Britain. The second classification shows the broad movement mechanisms by which material moves downslope. According to this classification less than half of the sites exhibit more than one type of mass movement, but a few exhibit more than two types. There is some correlation with the areal extent of a site and the number of types present, but this is not always the case. For example, the Axmouth-Lyme Regis GCR site runs along about 10 km of coastline, and exhibits six of Hutchinson's (1968a) types, while Quiraing, part of the Trotternish Escarpment GCR site, also a very large site, exhibits just one type. Most sites of small areal extent, however, exhibit a single type of mass movement. Rotational slips (groups 6a and 6bi) are the most common; the character of this type is discussed in further detail below.

(Table 1.5) The mass-movement GCR sites described in the present volume; style and type are according to the World Landslide Inventory

(WP/WLI 1993), classifications are according to Hutchinson (1968a) and (1988) — described in (Table 1.1)a,b.

(Table 1.6) The sites described in the present volume classified by geological age and by WLI mass-movement type:

(PC = Precambrian–Cambrian; Si = Silurian; De = Devonian; Ca = Carboniferous; Ju = Jurassic; Cr = Cretaceous; Eo = London Clay; PI = Pleistocene; fa = fall; to = Topple; sl = slide; sp = spread; fl = flow; * = sites which display cambering and valley-bulging).

Representativeness

Since 1980 a focusing of GCR objectives has taken place, whereby 'representativeness' is a term now used to encapsulate many of the 18 selection criteria recommended in 1992 (Gordon, 1992). At the time of the original selection process (1980s), GCR sites were not selected on the basis of their ability to represent mass movements in different geological forma tions or areas of the country, but rather to create an inventory of the most important mass-movement sites in Great Britain by mass-movement type. In reconsidering the Mass-Movements GCR Block in the light of the more focused objectives in the late 1990s (when the present volume was commissioned), sites were reconsidered against a scheme of stratigraphical and, thereby, areal representativeness (compare with the US system of geological site selection for conservation; Cooper, 1985). This re-focusing has brought about a change to the approach to the present mass-movement GCR volume, such that the text is divided into chapters on the basis of stratigraphy (age of the geological strata in which the mass movements occur; (Figure 1.9)).

'Representativeness' involves the notion of what is typical, or 'archetypal', but it is important to note that 'atypical' or 'exceptional' sites may provide insights into the nature of 'type' examples, and this is also a criterion for the GCR.

Revision of the GCR in the future

Mass-movements studies, like any other science, are ever-developing, with new discoveries being made, and existing models being subject to continual testing and modification as new data come to light. Increased or hitherto unrecognized significance may be seen in new sites. Therefore, it is possible that further sites worthy of conservation will be identified in future years for the study of mass movements in Britain, as research continues. However, it must be stressed that the GCR is intended to be a minimalist scheme, with the selection for the GCR of only the best and most representative example of a geological feature, rather than the selection of a series of sites showing closely analogous features.

Legal protection of GCR sites

V.J. May and N.V. Ellis

The list of GCR sites has been used as a basis for establishing Earth science Sites of Special Scientific Interest (SSSIs), protected under the Wildlife and Countryside Act 1981 (as amended) by the statutory nature conservation agencies (the Countryside Council for Wales, Natural England and Scottish Natural Heritage).

The SSSI designation is the main protection measure in the UK for sites of importance to conservation because of the wildlife they support, or because of the geological and geomorphological features that are found there. About 8% of the total land area of Britain is designated as SSSIs. Well over half of the SSSIs, by area, are internationally important for a particular conservation interest and are additionally protected through international designations and agreements.

About one third of the SSSIs have a geological/ geomorphological component that constitutes at least part of the 'special interest'. Although some SSSIs are designated solely because of the importance to wildlife conservation, there are many others that have both such features and geological/geomorphological features of 'special interest'. Furthermore, there are localities that, regardless of their importance to wildlife conservation, are conserved as SSSIs solely on account of their importance to geological or geomorphological studies.

Therefore, many SSSIs are composite, with site boundaries drawn from a 'mosaic' of one or more GCR sites and wildlife 'special interest' areas; such sites may be heterogeneous in character, in that different constituent parts may be important for different features.

Many of the SSSIs that are designated solely because of their Earth science features have interesting wildlife and habitat features, underlining the inextricable links between habitat, biodiversity and the underlying geology and geomorphology.

It is evident from some of the individual site reports in this volume, describing sites in coastal locations, that the conservation interest of the geomorphological features is likely to be affected by shoreline management activities outside of the site itself, especially where the GCR sites lie within large sediment-transport cells. A number of the sites have landslide toes which extend below low-water mark of spring tides. However, since SSSI notification of GCR sites presently extends to mean low-water mark in England and Wales and low-water mark of spring tides in Scotland, there is no statutory protection of these landslide toes below low-water mark, unless they are co-incidentally part of some other conservation designation (e.g. Special Protection Areas or Special Areas of Conservation — see below).

International measures

Presently, there is no formal international conservation convention or designation for geological/geomorphological sites below the level of the 'World Heritage Convention' (the 'Convention concerning the Protection of the World Cultural and Natural Heritage'). World Heritage Sites are declared by the United Nations Educational, Scientific and Cultural Organisation (UNESCO). The objective of the World Heritage Convention is the protection of natural and cultural sites of global significance. Many of the British World Heritage Sites are 'cultural' in aspect, but the Giant's Causeway in Northern Ireland and the Dorset and East Devon Coast ('the Jurassic Coast') are inscribed because of their importance to the Earth sciences as part of the 'natural heritage' — the Dorset and East Devon Coast World Heritage Site is of particular relevance here insofar as it was the outstanding geology and coastal geomorphology (including sites described in this volume and other sites described in the Coastal Geomorphology of Great Britain GCR volume (May and Hansom, 2003) that include mass-movement phenomena).

In contrast to the Earth sciences, there are many other formal international conventions — particularly at a European level — concerning the conservation of wildlife and habitat. Of course, many sites that are formally recognized internationally for their contribution to wildlife conservation are underpinned by their geological/ geomorphological character, but this fact is only implicit in such designations. Nevertheless, some of the sites described in the present volume are not only geomorphological SSSIs, but also habitat sites recognized as being internationally important. These areas are thus afforded further protection by international designations above the provisions of the SSSI system.

Special Areas of Conservation (SACs)

Of special relevance to the present volume are those coastal and mountain habitats that are dependent upon coastal or mountain geomorphology and are conserved as Special Areas of Conservation (SACs). In 1992 the European Community adopted Council Directive 92/43/EEC on the conservation of natural habitats and of wild fauna and flora, commonly known as the 'Habitats Directive'. This is an important piece of supranational legislation for wildlife conservation under which a European network of sites is selected, designated and protected. The aim is to help conserve the 169 habitat types and 623 species identified in Annexes I and II of the Directive.

Special Protection Areas (SPAs)

Special Protection Areas are strictly protected sites classified in accordance with Article 4 of the EC Directive on the conservation of wild birds (79/409/EEC), also known as the 'Birds Directive', which came into force in April 1979. They are classified for rare and vulnerable birds, listed in Annex I to the Birds Directive, and for regularly occurring migratory species.

Although SACs and SPAS are identified for the conservation importance of their biological features, individually or collectively, many also include scientifically important geomorphological features.

GCR site selection in conclusion

It is clear from the foregoing that many factors have been involved in selecting and protecting the sites described in this volume. Sites rarely fall neatly into one category or another; normally they have attributes and characteristics that satisfy a range of the GCR guidelines and preferential weightings (Ellis et al., 1996). A full appreciation of the reasons for the selection of individual sites cannot be gained from these few paragraphs. The full justification and arguments behind the selection of particular sites are only explained satisfactorily by the site accounts given in the subsequent chapters of the present volume.

Organization of the mass-movements GCR volume

The original plan for this volume was to divide it into chapters on the basis of mass-movement type. Thus, there would be a chapter on rotational slide sites, another on bedding-plane controlled slide sites, and so on. It was quickly realized that this would fail to represent adequately the network of GCR sites actually selected. In particular it separated some sites, which, when placed together, illustrated very well the variety of mass movements found in particular areas of the country, for example the southern Pennines. Since most of the sites illustrate complex landslides involving several types of failure, rather than single mechanisms, classification would be difficult.

However, a succession of chapters, some of which were based on mass-movement type, while others were based on regional considerations, gave a disorganized impression. Accordingly the present stratigraphical arrangement was adopted. This is still less than ideal. While it works well in highlighting the main mass-movement producing systems in Great Britain: Carboniferous, Jurassic and Cretaceous strata (which together account for 75% of the landslides identified 'in the DoE survey; Jones and Lee, 1994), it is less successful for sites in other geological systems. Since all of the mass movements in Great Britain represented by the sites described in the present volume have taken place in Quaternary times, the relevance of the age of the rocks in which they have taken place is indirect. More significant factors include the attitude of bedding, the frequency of jointing, and above all the succession of lithological types cropping out down a slope or a coastal cliff In particular, and this is the key to the prolific numbers of landslides in the Carboniferous, Jurassic and Cretaceous age rocks, is the presence of soft, 'incompetent' strata cropping out downslope or 'down cliff' from hard, jointed, 'competent' strata. An attempt to develop an ad-hoc order of presentation for the present volume was based on characteristics of physiography and geological succession at the selected sites. However, the arrangement of chapters by geological system for the purposes of publishing the accounts has been retained (Figure 1.9).

Comments on some general aspects of the sites selected

In addition to this introductory chapter in which general matters of relevance to the whole book are discussed, each of the following chapters has an introductory section in which geomorphological principles pertinent to the sites described in that chapter are discussed. However, some issues are described in the following text, which are of relevance to more than one of the chapters of site descriptions.

Movement

Mass-movement sites have in common that they represent the results of mass movements, i.e. movements that have already taken place; in other words, in only some of the selected GCR sites has the actual occurrence of movement been detected and recorded as it occurred. Movement may be detected in two main ways: measurement and eyewitness accounts. Measurement is generally carried out by identifying a fixed point (or points) on the ground surface and marking it/them with wooden or metal pegs. Its precise position is then surveyed, generally by triangulation from two locations whose positions are already known, by marking on a recent aerial photograph of the site, or using GPS techniques. After a period of time, perhaps one year later, the process is repeated using the same survey points (and taking new aerial photographs, GPS or laser-measurement data). A difference in the position of the marked point will indicate that mass movement has indeed taken place, and data can be recorded about the distance and direction of the movement. The problem with measurement of this type is that it is only worthwhile at a location where movement may be expected to take place over the surveying period, for example a location where movement is believed to have taken place in the previous year. However, the technique has been successfully used at East Pentwyn, Bourneville, Ironbridge, Mam Tor, St Mary's Bay, Black Ven, Stonebarrow, Folkestone Warren, St Catherine's Point, and the north coast of the Isle of Sheppey.

Eyewitness accounts exist for the 1839 movement of part of the Axmouth–Lyme Regis coast, now a National Nature Reserve. Eyewitness accounts were also collected by John Wesley, the Methodist preacher, of the collapse of Whitestone Cliff, North Yorkshire, in 1755 (reproduced in Jones and Lee, 1994; see also Cooper, 1997). Such accounts also exist for movements at, for example, Black Ven in Dorset, Robin Hood's Bay in North Yorkshire, and on the north coast of the Isle of Sheppey in Kent. Black Ven in Dorset is an important site where one can rely on seeing mudslides in motion, if visiting at the right time of year and after a suitable spell of wet weather. Black Ven has been intensively studied, with a complete record of movements for over 50 years. Stonebarrow, the next cliff to the east of Black Ven, has displacement and pore-pressure records for three years in the late 1960s, and the slides at Lyme Regis are currently heavily monitored. Many other records for short periods are associated with sites which require engineering stabilization works (e.g. Mam Tor). The fact remains, however, that many of the mass-movement GCR sites are only known in terms of simple morphological or geological descriptions.

Mudflows, mudslides and earthflows

Mudflows are generally taken to be rheological flows of material that consist predominantly of clay-sized particles, under the influence of gravity, and sufficiently wetted for the moisture content to be above the 'Plastic Limit'.

Mudslides are taken to be similar to mud-flows, except that they experience shearing at the contact with adjacent solid material. This zone of shearing is usually as sharp as a knife cut, with a 'scraped off' soft layer immediately above. The shear surface will be polished and striated. Deep-seated slides and extrusion layers may have a thicker zone of displacement. Mudslides can form within mudslides as they dry out, but still they are bounded by separate, clear shears (Brunsden, 1984).

This distinction largely became acknowledged with the publication of an important paper by Hutchinson and Bhandari (1971), in which it was explicitly recognized that many of the mass movements previously described as mudflows actually advance by sliding on discrete boundary shear surfaces, and that such mass movements are better termed 'mudslides', a term used by Fleming (1978) and by Cailleux and Tricart (1950). It was demonstrated that very often the surging forward of a 'mudflow' was caused not by flowage, but by undrained loading of its rearward parts, the whole mass moving down-slope by sliding. However, the term 'mudflow' is still valid for very fine-grained flows, but it is also an old term for mudslides. Hutchinson's 1968a 'climatic mudflows' (see (Table 1.1)a above) are now mudslides (Brunsden, 1984).

In the World Landslide Inventory (WP/WLI 1993) classification, the American usage 'Earth-flow' is preferred. Buma and van Asch state in Landslide Recognition (Dikau et al., 1996) that 'the American usage 'earthflow' is replaced in European literature by 'mudslide". However, 'earthflow' is used by Skempton et al. (1989) in describing part of the landslide at Mam Tor (see Chapter 5). Varnes (1978), the principal American source on such matters, does not endorse this one-to-one correspondence in terminology. Stating that earthflows range in water content from above saturation to essentially dry, he places mudflows at the wet end of the scale, as 'soupy end members of the family of predominantly fine-grained earthflows'. This neglects the important observation that the 'stiffer' forms slide on discrete surfaces.

Undrained loading

Hutchinson and Bhandari (1971) provided an expanded account of a suggestion made by Hutchinson (1970) which applies to many mudslides and also to a variety of other types of mass movement. They observed that many 'mudflows' were advancing downslope by shearing on slopes that were of considerably lower angle than the slope of limiting equilibrium for residual strength on the sliding surface and groundwater co-incident with and flowing parallel to the slope surface. For example, with slopes at Bouldnor, Isle of Wight that have residual strength of c_r' = 0, ø_r' = 13.5° (where c_r' is residual cohesion and ø_r' is angle of internal friction), it was shown using infinite slope analysis (Skempton and Delory, 1957) that the lowest slope angle at which sliding could occur is 6.1°. Measurement of these slopes showed that they stand at angles as low as 3.9° (Hutchinson and Bhandari, 1971). They suggested that the sliding is brought about by the virtually undrained loading of the headward parts of the mudslides by debris discharged from steeper slopes to the rear. This undrained loading develops a forward thrust in the rear part of the mudslide, where the basal slip-surface is inclined fairly steeply downwards, giving rise to shearing movements on very low angle slopes (Figure 1.10), even at slopes of zero or negative inclination for short distances (Hutchinson and Bhandari, 1971).

Collapse of caprocks

There is a group of mass movements, generally characterized by a hard but possibly jointed caprock, which does not have to be thin and can be several tens of metres in thickness, overlying a stratum or series of strata characterized by 'incompetence', the inability to support the overlying 'competent' caprock at locations where erosion has cut down to expose the incompetent strata. This can lead, according to local circumstances, to one or more of a variety of recognized mass-movement types, in the terms of Hutchinson (1988) rebound associated with naturally eroded valleys, post-failure creep, and complex failures of types (1) cambering and valley bulging, and (2) block-type slope movements.

This phenomenon has been more widely accepted in continental Europe and other parts of the world than in Great Britain ' (Brunsden, 1996a). Possibly, British workers, who naturally are those who have been most closely concerned with British mass movements, have been too circumspect. Why invoke a thick mobile stratum when a thin one will do? This shows confusion between theory and verification. In a highly empirical subject like geology such theorizing must give way to evidence that shows nature to be more complex than expected.

Non-circular failure surfaces

Some failures take place over a surface which, when seen in section, has the form of an arc of a circle. A rotational slip over such a surface results in the slipped mass tilting backwards, and the form of the surface enables slipping to happen without the slipped mass breaking up. This observation has been used by geotechnical engineers to provide a simple method of analysing such 'circular failure', using 'circular failure charts' (see, for example, Hoek and Bray, 1977). This, in turn, has led to the expectation that many failure surfaces will have the form of a circular arc. Thus, many slipped masses which have rotated backwards are assumed to have rotated on a circular arc.

That this perception has been recognized as over-simple is illustrated by one of the differences between Hutchinson's 1968a and 1988 mass-movement classifications (Table 1.1)a and (Table 1.1)b. The term listric (spoon-shaped), used in (Table 1.1)b, refers to a surface that is at all points concave upwards, but of which the radius of curvature decreases downslope. This naturally causes the slipped mass to crack and break up. A further point tending to make circular failures rather unusual is that in sedimentary rocks, at least, the sedimentary sequence is rarely massive enough to be effectively anisotropic with respect to physical properties. As a result, whenever a failure surface meets a pre-existing plane of weakness, it tends to follow it, whether it be a fault, a joint or a bedding plane. An important result of this is that, in many cases, the failure plane may be a non-circular concave-upwards curve beneath the upper parts of a landslip, but follows a sub-horizontal planar bedding beneath the downslope parts (this argument is from Varnes (1978), although Barton (1984) traces it to Taylor (1948); see (Figure 1.11)).

Rib and Liang (1978) point out that down-slope decrease in the curvature of the failure plane produces tension and ultimate failure in the slump block owing to lack of support on its uphill side. This can lead to the formation of a graben in the rear of the slope (Figure 1.11). However, Barton (1984) has observed, from the opportunities that exist for the examination of 'rotational' slips in cross-section, that often the only concave-upwards segment of a slip-surface is of small radius of curvature, at the foot of a straight, steeply dipping segment, and grading into the angle of the bedding on its downslope side (Figure 1.12) and (Figure 1.13). He goes so far as to suggest that this is such a common observation worldwide that it should be 'regarded as the norm and such a surface should be assumed until, and unless, definite evidence to the contrary is obtained'. This conclusion is amply borne out by the mass-movement sites selected for the GCR. Further, although this is not mentioned in its accompanying text, the diagram illustrating a 'single' landslide in The Multilingual Landslide Glossary (Figure 1.8) shows a failure surface of this type.

General characteristics of GCR site descriptions

The length and detail of each site description herein has been determined by the volume of research that has been published on the site.

Generally, the most significant sites in terms of the development of understanding of mass movements in Great Britain are those that have received the most detailed study, often over a long period of time, and over which contending views may have developed. On the other hand, some sites have been selected about which very little has been written, but which nonetheless exhibit features of special interest. In these cases the text concentrates on general description rather than detailed scientific explanation. It is hoped that this will provide an incentive, justification and/or rationale for further research.

Overall, the site descriptions vary considerably in length, detail, and degree of illustration. To have imposed a rigid uniformity on the descriptions would have failed to give an accurate impression of the variety of mass-movement sites to be found in Great Britain, and would have failed to do justice to the most intensively studied sites, those with innovative and/or enterprising methods of study, and those with the longest history of study.

Consideration was given to providing each site description with a stereopair of aerial photographs, so that the physiographical expression of mass movement at each site may be illustrated. However, there is a risk that at some sites this could lead to an inappropriate concentration on the physiographical aspects of the site and not their scientific causes/ importance per se. Also, woodland or forest vegetation tends to obscure or smooth over such features as viewed from above.

Many of the site descriptions are provided with cross-sections. Where they are not, this is because no reasonably accurate cross-section has been published. However, some of the sites are illustrated by slope profiles measured by the present author (RGC). In all cases these were measured in successive 1.5 m ground lengths using a slope pantometer (Pitty, 1966). They run directly downslope, and are orthogonal to the contours. In order to avoid this" orthogonal line appearing as a curve in plan, locations for measurement were selected where the contours were roughly parallel to each other. All profiles were originally plotted at a scale of 1:400, and are drawn without vertical exaggeration.

References