Folkestone Warren, Kent

[TR 243 375]–[TR 268 385]

Introduction

(a) General

The principal reason for the detail in which Folkestone Warren has been studied has been that since 1844 the 3 km-long unstable area has been traversed by the main railway line from Folkestone to Dover (Figure 7.2). The photographs of the twisted railway tracks, one with a train on them, taken after the great landslip of December 1915, have become justly famous (Figure 7.3). Detailed investigations have been carried out by the Southern Railway and later by British Railways, and more recently by workers from a number of academic institutions and civil engineering consultancies.

The area of landslides is backed by the 'High Cliff', a 30 m-high Chalk cliff standing at about 55°, consisting of a succession of broad, irregularly spaced buttresses. In plan (Figure 7.4) and (Figure 7.5), the rear scarp of the Folkestone Warren has a generally arcuate form, concave to seaward, with the degree of concavity increasing towards the west. It is made up chiefly of three en-echelon sets of essentially joint-controlled faces (Hutchinson et al., 1980), which trend at 67°, 47° and 109° (Figure 7.4). These give the rear scarp in detail a rather saw-toothed appearance. Its general form reflects the tendency of the faces at 47° to become increasingly predominant over those at 67° towards the west. The upper part of the cliff is vegetated, but much of its foot was freshened by the movements of 1915. The 300–400 m-wide area of slips between the High Cliff and the coastline is known as the 'Undercliff'. The multiple form of the Warren slips is best seen at the western end, where the rear scarps of slips of the 1930s and 1940s parallel both the associated rear scarp of the 1915 slip and the High Cliff above the ancient slide block of 'Little Switzerland' (Figure 7.6). Elsewhere the slide blocks are generally obscured by a mantle of Chalk debris derived from numerous falls from the High Cliff. There is a break in the talus at the foot of the High Cliff. This was produced by the 1915 slip. It can still be traced except where masked by later Chalk falls. The seaward edge of the Undercliff consists of sea cliffs up to 15 m high, exhibiting a great thickness of Chalk fall debris overlying the slip deposits.

(b) Stratigraphy

The arrangement of the undisturbed strata in the immediate vicinity of Folkestone Warren was described by Osman (1917). The cliffs are formed by the truncation of the scarp of the North Downs by the sea, and consist of Middle and Lower Chalk (Upper Cretaceous), overlying Gault Clay and the Folkestone Beds division of the Lower Greensand (Lower Cretaceous). All of the strata dip at about 1° in a direction between north-east and NNE. The Folkestone Beds formation is about 18 m thick and consists of coarse-grained yellowish greensands with bands of calcareous and glauconitic sandstone (Gallois, 1965). The junction between the Folkestone Beds and the overlying Gault Clay is at the top of the 'Sulphur Band', a bed of phosphatic nodules (Smart et al., 1966).

The Gault Clay consists of hard, over-consolidated, fissured and jointed clays. It is between 40 m and 50 m thick. Thirteen lithological subdivisions have been recognized within it (Jukes-Brown, 1900). These possess large variations in physical properties, for example in the liquid limit — in the upper part of the Gault Clay, liquid limit values lie between 80% and 120%, falling to 60%-70% in the middle part, and rising again to 90%–110% in the lower part, with a rapid fall to about 70% immediately above the Sulphur Band (Toms, 1953; Muir Wood, 1955a).

The lowest bed of the Lower Chalk is the Chloritic Marl, a relatively impermeable formation about 27 m thick, which can have as much as 50% argillaceous and arenaceous matter (Gallois, 1965). The overlying Grey Chalk, 24 m thick, becomes more pure, massive and blocky upwards into the overlying well-jointed White Chalk, 18 m thick and itself overlain by a 2 m-thick layer of plenus marls. The super-incumbent Middle Chalk has at its base the Melbourn Rock, a band of nodular, gritty, yellowish-white Chalk about 12 m thick. This passes upward into a fine white Chalk in massive, well-jointed and highly pervious beds. At Folkestone Warren these beds extend up to the level of the presumed Pliocene or Plio–Pleistocene platform (Jones, D.K.C. 1980) which defines the top of the cliffs. Detailed descriptions of the Middle Chalk at Horse's Head, the Gault Clay west of Horse's Head, and Glauconitic Marl and Chalk Marl on the foreshore east of Horse's Head, are provided by Gale (1987).

(c) History of landsliding

The first recorded landslide at Folkestone Warren took place in 1716. Since then failures have recurred at frequent intervals. As noted by Muir Wood (1955b), slipping has been more frequent in the western part of the Warren. Since 1844 there have been 11 known deep-seated landslides. All have taken place within the December to March period, which is when groundwater levels are highest. Moreover, the three largest known failures have occurred in the first third of this period. This suggests that the landslides were brought about by particularly high seasonal peaks of groundwater pressure in the slipped masses. In the case of the 1915 slip at least, there is evidence that it took place, as might be expected, near the time of low tide.

Studies at The Roughs (Brunsden et al., 1996a), 12 km to the west of Folkestone Warren, confirm this. Geotechnical data from investigations at the site have been plotted on every cell of a slope map derived from a digital terrain model (DTM) of the site, to produce a factor of safety map of the Hythe Beds escarpment. Combining this with 5% and 8% perturbations of the current climatic trend extrapolated to the years 2030 and 2050 respectively, shows that many currently dominant areas of the escarpment are vulnerable to a small rise in water levels. Statistical analysis of archive data confirms this conclusion. Since the Weald Clay escarpment is adjacent to Folkestone Warren (the next feature along the coast to the east), it is likely that this conclusion applies there too.

The incidence of Chalk falls appears to be little influenced by the seasonal variations in groundwater level. This no doubt reflects the fact that the falls involve chiefly the body of Chalk situated above the highest position of the groundwater table in the slipped masses and adjacent Chalk. The boreholes of Trenter and Warren (1996) reveal three en-echelon slips that commence north of the railway, trend eastwards and swing south-east to the beach, surfacing on the foreshore at the Warren Point, Horse's Head and Warren East End locations. Measurements made at the toes of each of these slips show that in each case the rate of movement increased eastward to a maximum where the slip crossed the foreshore. On the adjoining landward side, movements were markedly smaller.

There has also been sliding or falling of large masses of Chalk from the High Cliff at the rear of Folkestone Warren. Chalk falls from the rear of the Warren are commonly preceded by slight, chiefly downward, movements known as 'sets'. These affect the Chalk behind the High Cliff for distances of probably up to 20 m, and may also involve the underlying Gault Clay (Toms, 1953). A subsidence of as much as 1.5 m has been recorded, but movement is usually much less than this. It is noteworthy that the great majority of the three dozen recorded failures have consisted of renewals of movement in the slipped masses that form the Undercliff and have in no case involved a general recession of the rear scarp of the landslips. Thus, although locally scarred by Chalk falls, or slightly shifted by sets, the High Cliff is a feature of considerable age.

The largest slip about which detailed information has been collected took place in 1915. Other notable movements occurred in 1937 and 1940 (see (Table 7.1) and (Table 7.2). The 1937 landslip (Toms, 1946; Hutchinson, 1969) was more than 900 m wide, and affected the whole of the slope seawards of the railway line (Figure 7.7). Upheaval of the foreshore took place. Seaward movement varied from about 27 m in the western part of the slip to zero at Warren Halt. The 1940 landslip (Toms, 1953; Muir Wood, 1955b; Hutchinson, 1969) took place in about 6 ha of the Warren, with a length along the coast approaching 700 m. Movements were slight and gradual, beginning in 1940 and continuing episodically to 1947, amounting to 1.5 m horizontally at most. Its essential features are shown on two cross-sections (Figure 7.8).

Since 1936–1940, apart from a slight renewal of movement of the 1940 slip in 1947, no major movements have occurred. The improved stability of the Warren since 1915 is probably the cumulative result of coast protection measures, drainage works in the slipped masses and the extensive weighting of the toe described by Viner-Brady (1955).

Description

(a) Hydrology

Investigations in 1948–1950 left little doubt that most of the groundwater in the slipped masses derives from the aquifer provided immediately inland by the Chalk. Using data from boreholes in the Warren, Muir Wood (1955b) has drawn contours on piezometric levels in the slipped masses. Monthly observations from 1953 onwards give a good indication of the seasonal fluctuation in piezometric levels, which is of the order of 3–9 m (Hutchinson, 1969) (Figure 7.9). Comparison of hydrological measurements within the Warren with records for a nearby Chalk well shows that an intimate connection exists between the groundwater bodies in the slipped masses and the adjacent in-situ Chalk (Muir Wood, 1955b).

(Table 7.1) Folkestone Warren: summary of the average values of ø_r' (°), σ_n' and s in the Gault Clay at failure in the 1940, 1937 and 1915 landslips. The original pre-metric data have been used. After Hutchinson (1969).

(Table 7.2) Results of stability analyses. After Hutchinson (1969).

At The Roughs, analysis using the 5% and 8% perturbations of the current climatic trend mentioned (Brunsden et ed., 1996a), showed that only a small rise in water levels would bring groundwater to ground surface level over large areas of the Hythe Beds escarpment. Again, since this is so close to Folkestone Warren, it is likely that the same is true there also.

Available evidence suggests that the seasonal variation in groundwater levels in the Folkestone Beds beneath the Warren is small. At the Warren it is likely that the levels in this confined aquifer fluctuate slightly in response to, but lag somewhat behind, the tidal variations in sea level (Hutchinson, 1969).

(b) Coastal factors

Hutchinson et al. (1980) have discussed the predominant eastward littoral drift on this part of the coast (Figure 7.10), and the way that interference with this drift in the Folkestone area tends to deplete the beaches downdrift and hence stimulate landsliding in Folkestone Warren. Major interference dates from the construction of the masonry harbour at Folkestone in the first decade of the 19th century. Construction began in about 1807. The West Pier was completed by 1810 and the Haven by 1820. However, by 1830 the harbour mouth was completely choked by sand and shingle. From 1842 onwards, successive pier extensions were carried out in order to produce shingle-free, deep-water berths (Figure 7.11). In 1861–1863 the Promenade Pier was built, initially of timber but stone-cladded and extended in 1881–83 when it was renamed the New Pier. A further extension to this was made in 1897–1915. Accelerated coastal erosion at the Warren was commented upon by Drew (1864) and Osman (1917), as well as by many local authors. Hutchinson et al. (1980) conclude that these works led to a progressive increase of landsliding in the Warren, which culminated in the great slide of 1915.

(c) Geotechnical investigations

A few boreholes were put down in Folkestone Warren in 1916, shortly after the major movement, but they yielded little information. The first co-ordinated geotechnical investigation was made by the Southern Railway in 1938–1939, following the landslip movements in 1936 and 1937. The investigation was concentrated towards the western end of the unstable area, and was described by Toms (1946), who concluded that the slides are large-scale slumps of Chalk over underlying Gault Clay, on non-circular slip-surfaces. More extensive investigations made between 1948 and 1950, reported in detail by Toms (1953) and Muir Wood (1955b) led these authors to concur in this interpretation.

Muir Wood (1955b, 1971) noted that the rotational slips penetrate to the base of the Gault Clay and that failure is largely confined to a plastic sheet of clay immediately overlying the 'Sulphur Band'. Thus, testing of the strength of the materials involved in the failures has concentrated on the Gault Clay. Toms (1946) measured its unconfined drained shear strength, but as the landslips considered have all involved renewals of movement upon pre-existing slip-surfaces where shear displacements of tens of metres have taken place, the shear strength mobilized at failure can be taken to be residual. The residual strength of the Gault Clay, taking samples of both high and low liquid limit, has been determined by Hutchinson (1969) (Figure 7.12)a and, using more refined ring-shear techniques, by Hutchinson et al. (1980) ((Figure 7.12)b). A sample from the high liquid limit zone near the base of the stratum has a residual angle of shearing resistance, ø_r', of 12° (Hutchinson, 1969). For a sample from the lower liquid limit material forming the middle part of the stratum, he obtained a value of 19° for ø_r'. However, using ring-shear tests, Hutchinson et al. (1980) found 12° for the low liquid limit Gault Clay and 7° for the high liquid limit material (Figure 7.13). These values are up to 7° lower than those obtained by Hutchinson (1969) in cut-plane direct shear tests on similar material at lower normal effective stresses. In relation to the Chalk falls, Hutchinson also measured the residual strength of a sample of the Middle Chalk, obtaining c_r' = 0, ø_r' = 35°.

The Folkestone Warren landslips have been the subject of a large number of stability analyses. The 1937 landslip was analysed in terms of total stresses and using a rotational landslide failure model, by Toms (1946) and Skempton (1946). Similar analysis of the 1915 and 1940 landslips was carried out by Muir Wood (1955b). The 1915, 1937 and 1940 landslips have been analysed in terms of effective stresses by Hutchinson (1969) and Hutchinson et al. (1980), an approach more appropriate for longterm problems. In contrast to the earlier work of Toms (1946, 1953) they employed a non-circular, multiple failure model. Hutchinson (1969) found that the average strengths mobilized on the non-circular failure surfaces in the Gault Clay approximate to the residual, and are bounded by the envelopes c_r' = 0, ø_r' = 13.9° and c_r' = 0, ø_r' = 16.6°. However, Hutchinson et al. (1980) found that in the range of average normal effective stress levels in the Warren (about 200–800 kN m⁻³), the values of 9); indicated as likely by stability analyses are more probably in the range 7.5° to 15°. Clearly, these values tend to be higher than those derived from ring-shear tests. Trenter and Warren (1996) made residual effective shear strength determinations on Gault Clay samples using the Bromhead ring-shear device and the reversal shear-box, obtaining results in good agreement, with an average ø_r' value of 9.5° (Figure 7.14). Back-analyses for two of their cross-sections (Warren Halt and Horse's Head) produced ø_r'values averaging 10.7°. They see the back-analysed results as being reasonably close to the average measured value. They are also close to the field residual strength line obtained by Skempton et al. (1989) at Mam Tor (Figure 7.13).

The value of 10.7° was obtained without including the result from the Horse's Head slip 2, which showed differences between back-analysed and measured o_r' values. These are explainable in part as resulting from a proposed 'hinging' mechanism for these seaward landslides: 'as more movement occurs along pre-existing slip planes in the Gault Clay to the east, caused by a rise in the groundwater table in the chalk or chalk rubble (or, before the protection works, by erosion), some first-time movement will be provoked about the hinge further west' (Trenter and Warren, 1996, p. 618).

Interpretation

(a) Classifications of slide types

Two main views have emerged regarding the types of slides present at Folkestone Warren. Hutchinson (1968b, 1969; Hutchinson et al., 1980) followed earlier workers in taking the view that most major movements at Folkestone Warren are rotational albeit controlled by a planar basal, bedding-plane, surface. He divided the recorded landslides into three main types. The largest of these ('Type M', for multiple rotational) involve a renewal of movement in virtually the whole of the landslips which form the Undercliff, and result in large seaward displacements of the railway lines. Smaller features of rotational character (Type R', for single rotational) comprise a renewal of movement only in the slip masses in the vicinity of the sea cliff. The remaining type is the sliding or falling of large masses of Chalk from the High Cliff at the rear of the Warren (Figure 7.15).

The second and more recent view is that of Trenter and Warren (1996) who gave more emphasis to the effect of the planar bedding control. Based on borehole observations, it provides an alternative two-fold classification of the mechanisms of the slips, which does not coincide with Hutchinson's (1969) Type M and Type R. Trenter and Warren's 'Slip 1' type consists of large translational slips extending from the High Cliff to the sea, with a failure surface passing through the basal Gault Clay, immediately above the Sulphur Band. Their 'Slip 2' type corresponds to smaller features, on circular failure surfaces at the east end of the Warren but compound at the west with a failure surface passing through substantial quantities of slipped and broken chalk.

(b) Retrogression mechanism

A possible mechanism for the retrogression of the rear scarp of the landslip has been suggested by Hutchinson (1969) (Figure 7.16). He recognized that the existence of large horizontal stresses in over-consolidated plastic clays is well attested, as is the lateral expansion that such deposits exhibit under reduction of their lateral support. At Folkestone Warren the over-consolidated Gault Clay lies between two more rigid strata, the Folkestone Beds below and the Chalk above. The available field evidence suggests that the lateral expansion of the seaward parts of the Gault Clay, resulting from the reduction of their side support by marine erosion and landsliding, will have been accompanied by the generation landwards of a shear surface of residual strength situated at or near the base of the Gault Clay. This in turn suggests that the failure of the 'sets' does not take place at peak strength. A possible mechanism for the progressive failure involved in the generation of such a shear surface is given by Bjerrum (1966). These 'first-time' failures take place at 'residual' strength and not at 'peak'. This is also the mechanism proposed for the Castle Hill landslide at the portal of the Channel Tunnel. A further result of the seaward expansion of the Gault Clay is that the Chalk caprock, acting as a sensitive movement indicator, will have been thrown into tension with the consequent opening-up of those vertical joints behind but close to the High Cliff. Widened joints locally known as 'vents' have been found from time-to-time behind the cliff-top. These are generally filled by superficial deposits, for example Pliocene sands. It seems likely, in the absence of anticlinal flexure and/or cambering, that they were initiated by lateral extension of the hill.

The forces resisting the collapse of the Chalk mass isolated between the cliff-face and the most seaward vent will have been greatly diminished by these movements. If insufficient support is provided by the slipped masses to seaward, the block will fail and begin to subside. With increasing subsidence the curvature of the failure surface produces a back-tilt of the failing block (the apparent rotational failure) which will bring about a re-engagement of part of the irregular joint faces. Although the whole failure surface will now be at its residual strength, the stabilizing contribution from the re-engaged joint may be sufficient to effect a temporary cessation of movement. Such a mechanism is thought to be the explanation of the arrested failures in the High Cliff known as 'sets'. Final collapse of the block will generally coincide with the next reduction of support by the slipped masses. The cycle of retrogression may then be repeated on the next block to landward.

The major part of the lateral expansion in the Gault Clay, and therefore of the formation of the associated vents and basal shears, seems likely to have been roughly contemporaneous with the initiation of the present landslides. The effect of this can be expected to have led to relatively rapid, successive cycles of retrogression of the rear scarp. These will have proceeded until a situation similar to that of the present day was reached in which the support provided by the slipped masses is generally sufficient to prevent the total collapse of the Chalk forming the current rear scarp.

These ideas are broadly supported by Trenter and Warren (1996), whose borehole information shows the form and nature of the slipped masses comprising the Undercliff varying along the Warren's length (Figure 7.17), (Figure 7.18), (Figure 7.19), (Figure 7.20), (Figure 7.21). At the western end the High Cliff reaches its highest point at about 165 m above OD while the Undercliff is at its lowest at about 70 m above OD. However, at the eastern end, the High Cliff is lower at 135 m above OD, with the Undercliff at its highest, 90 m above OD. Trenter and Warren's (1996) boreholes show up to 45 m thickness of broken Chalk at the west end, while at the east end the mantle of broken Chalk is only 10–15 m thick. The boreholes also indicate larger amounts of slipped but intact Lower and Middle Chalk at the east end, the prime example being the Horse's Head, a prominent small hill, at Horsehead Point (Figure 7.18), (Figure 7.19), (Figure 7.20), (Figure 7.21). They consider these results to indicate that the landslides at the west end of Folkestone Warren are of much greater size and mobility than the rest, and suggest that this is due partly to more massive Chalk falls from the High Cliff behind, and partly to the consequent undrained loading of the Undercliff landslides. They also point out, following Hutchinson (1969) and Muir Wood (1994), that most of the Folkestone Warren slides have been confined to the shoreward part of the Undercliff and the foreshore. Only the slides of 1877, 1896 and 1915 penetrated as far as the High Cliff.

(c) The 1915 landslip

The 1915 failure is complicated by the fact that several large Chalk falls from the High Cliff were associated with it (Figure 7.22). As noted by Hutchinson (1969) this introduces uncertainties into stability analyses. An unusually large number of eyewitness accounts of the slide were collected. These tend to show that the renewal of movement commenced at the western end of the Warren and spread eastwards. This disturbance was followed by three associated failures of the rear scarp in a west to east order, which were therefore triggered by the slips, rather than initiating them (Hutchinson et al., 1980). However, the falls doubtless further stimulated the movements of the Undercliff. Field and historical evidence suggests that Chalk falls occur at the projecting corners of the individual 'saw-teeth' of the irregular rear scarp of the landslip area.

Hutchinson et al. (1980) also showed that in the year 1915 the rainfall recorded at Folkestone had the highest annual total since records had begun in 1868, 138% of the 1881–1915 average. The September–December 1915 total was the second highest since 1868 (158% of the 1881–1915 average), and the December 1915 total was the highest since 1868 (256% of the 1881–1915 average). They attribute the renewal of movement in the Undercliff to this unusually high rainfall, and to the erosion of the bay as intensified by pier construction. The low tide (two days before Spring tides) that preceded the first reported movements by about 1.25 hours, was probably the final trigger. The study at The Roughs (Brunsden et al., 1996a) indicates that, given only small perturbations to present climatic trends, similar conditions may become more common, and it may be presumed that this would be accompanied by similar slope movements.

In the main part of the 1915 slide, the displacement (as measured by the movement of the railway lines) was 10–20 m, increasing to a maximum of about 50 m at one point (Figure 7.22). In the central part of the Warren, the Undercliff is blanketed by flow slide debris from the 1915 fall. In a discussion of Hutchinson (1969), Muir Wood (1970) commented on the fact that the 1915 slip moved forward about 30 m, and asked how this could be reconciled with a renewal of movement on a pre-existing, residual slip-surface. Hutchinson et al. (1980) listed nine possible mechanisms for this, and in view of the large Chalk fall from the High Cliff that took place just after the start of the slow movement in 1915, concluded that sudden undrained loading (Hutchinson and Bhandari, 1971) of the rear of the slide by this rockfall, was the most likely explanation.

Casey (1955) described how the 1915 slide was accompanied by an upward heaving of the foreshore, involving a strip that extended between 140 m and 240 m out from the shoreline. In front of the central half of the main slip there was a second ridge of upheaved seabed farther seawards. The outer edge of this reached to between 260 m and 350 m from the former shoreline and it had a width of 80–90 m (Figure 7.23). The inner and outer zones of upheaved material were partly separated by a lagoon. Earlier large slides at Folkestone Warren were also accompanied by the development of long Gault Clay and Chalk ridges that formed on the foreshore. Trenter and Warren (1996) suggest that these ridges or reefs may have been due to seaward movement of their Slip 2 slides thrusting under the slipped masses lying beneath the foreshore, and so raising them. However, since Casey (1955) describes such ridge development as having taken place some time before the main movements of 1916 took place, they also consider it possible that prior to the onset of slipping there may have been some plastic flow of the Gault Clay, with the failed Gault Clay erupting at the foreshore.

(d) Contemporary movements

Trenter and Warren's (1996) triangulation and observations within drainage headings, supplemented by piezometer readings, have shown that an area in the vicinity of the 'Horse's Head', which is the most prominent sea cliff in the Warren, in the central part of the Undercliff, has been moving seawards since at least 1938. The area of greatest movement co-incides with the location of the debris tongue of the 1915 fall, and is also the area in which coastal erosion was at a maximum until the extension of the present sea wall across it. Piezometer results suggest that at least in the vicinity of the seaward edge of the Warren and possibly elsewhere, the slipped masses consist of blocks of Gault Clay of various sizes among Chalk blocks and debris which in general have become unloaded as the slips have developed (Hutchinson et al., 1980). The piezo-metric pressures within the Chalk and the smaller masses of Gault Clay had equalized with the long-term groundwater conditions, but negative excess porewater-pressures still existed within the larger blocks of slipped Gault Clay (Figure 7.24). As these swelled back to equilibrium the factor of safety on any slip-surfaces traversing these blocks steadily decreased.

(e) Age

It is unlikely that Folkestone Warren was initiated before the virtual completion of the Flandrian transgression. An immature soil profile situated towards the base of Chalk debris overlying a bluff of steeply back-tilted Middle Chalk, at the Horse's Head, has been dated by its contained fauna to Atlantic or Sub-Boreal age. As these sea cliffs are the oldest remaining part of the present Warren landslides, and the soil layer is also strongly back-tilted, it can be inferred that the main slipping movements affecting it have occurred since its formation, suggesting a Sub-Boreal date (5500–2500 BC) for the initiation of the present landslides (Hutchinson, 1968b). The weathered and vegetated appearance of those parts of the present High Cliff that have not suffered relatively recent Chalk falls suggests that its age of at least two centuries indicated by the historical records is on the low side.

Conclusions

At least before the construction of the sea wall and toe weighting, Folkestone Warren was in a state of dynamic equilibrium under the combined influences of its topography, hydrology, geology and its exposure to marine attack. This is reflected chiefly in the widening of the Undercliff towards the west, which compensates for the gradual reduction in passive resistance at the toe as the level of the Gault Clay rises.

Folkestone Warren is a mass-movement site of great importance, particularly due to the great detail in which, and timescale over which, it has been studied. Other large-scale Chalk slips on the south coast, for example the Undercliff at Ventnor (Isle of Wight), and White Nothe in Dorset, seem to be broadly similar in nature. The results of the studies on Folkestone Warren are quoted in many studies of mass movements still further afield. The site is fundamental in the development of understanding of both translational and rotational slips, and the relationships between them.

The early recognition of the clay-extrusion model is of fundamental importance and has since become important in the interpretation of the major slides at Castle Hill on the Channel Tunnel portal, and is being increasingly recognized as a fundamental failure mechanism. New research on this topic is in progress in many countries.

References