1. Introduction
Almost the entire length of this coastline is characterised by active cliff development, with adjoining beaches and shore platforms of variable length, height and width e.g. Photo 1. Contrasts in cliff height and morphology are the product of the outcrop pattern of Cretaceous rocks of varied lithology, and of controls imposed by geological structure. The western and eastern coastlines truncate the axis of the Isle of Wight monoclinal fold, such that the same sequence of outcrops is encountered; local variations in rock dip introduce spatial variations in outcrop width (Osborne White, 1921; Bird, 1997). The southern coastline is distinctive in that it is developed in debris that is the product of long-established slope instability (Photo 2) (Preece, 1980; 1987; Chandler and Hutchinson, 1984; Hutchinson, Brunsden and Lee, 1991; Hutchinson, 1991; Hutchinson and Bromhead, 2002). The landslides of this sector of the island's coast owe their fundamental character and impressive scale to rock lithology and succession, climatic history, hydrogeological controls both above and below mean sea-level, structural form, and wave climate. (Halcrow, 1997; Hutchinson, 1991; Rendel Geotechnics, 1995; Royal Haskoning, 2010).
A major new source of coastal data is from the Defra-funded National Network of Regional Coastal Monitoring Programmes. The Programmes consist of topographic beach surveys, nearshore bathymetry, aerial photography, lidar, coastal hydrodynamics (waves and tides) and terrestrial habitat mapping. Specifications for data collection are consistent for all regional programmes and the data and analysis reports are made freely available under the Open Government Licence from www.channelcoast.org.
In 2011 the Coastal Monitoring Programme completed an high resolution, 100% coverage swath bathymetry survey of the nearshore zone of the northern and southern coasts of the Isle of Wight, extending 1km offshore from the MLW. This southern survey area abutted a survey further offshore commissioned by Natural England, which completed the entire coverage of the extent of the marine Special Area of Conservation.
The Southeast Regional Coastal Monitoring Programme commenced in 2002. The Lead Authority is New Forest District Council, with data collection, analysis and reporting led by specialist teams at the Channel Coastal Observatory (CCO), Canterbury City Council and Adur and Worthing Councils. (See CCO Annual Survey Reports for further details).
1.1 Coastal Evolution
Variations in coastal orientation, wave exposure, relief and geological outcrops enable sub-division into three distinct behavioural units summarised as follows, partly based upon evaluations undertaken for Halcrow (2002):
Sandown Bay
1. Differential erosion of soft clay, shales and sandstones of the Wealden and Lower Cretaceous formations has formed an embayment anchored by headlands, controlled by the occurrence of more resistant Chalk to the north.
2. Erosion would have operated over at least the past 5,000 to 6,000 years, since rising postglacial sea-level has approached its present elevation. Extensive shore platforms provide evidence for long-term recession in outcrops of more resistant bedrock, and appear to extend well seawards of low water. It is thought that several kilometres of recession has occurred over a timescale in excess of ten millennia.
3. Erosion has released large quantities of predominantly sandy sediments creating the fine sandy beaches of the bay. Whilst some sediments have remained within the bay, most have been transported elsewhere. On the basis of mineralogy (Dyer, 1980; Algan, et al., 1994), it has been suggested that this material could have contributed to Ryde Sands, although other areas of potential accumulation also exist to the east of the bay.
4. Coastal recession has truncated a tributary of the E. Yar valley at Yaverland. Sediments migrating into this valley in the form of a barrier beach appears to have prevented more recent marine inundation and has preserved the regular planform of Sandown Bay.
The "Undercliff"
1. Almost the entire coast has developed within the debris of massive sequent landslides thought to have been initiated prior to Holocene rise in sea-level when their basal debris lobes, mantled by periglacial colluvial material, would have extended considerably further seaward than at present. The seabed between the modern shoreline and St. Catherine’s Deep, some 2km. offshore, is formed of planed-off landslide blocks and derived debris.
2. Late Holocene sea-level rise re-occupied the toes of these relic landslides eroding them and reworking their sediments to remove much of the lower slope deposits thereafter cutting into the displaced landslide blocks of the Undercliff itself.
3. The majority of sediments eroded were relatively non-resistant and rapidly removed by wave action. However, large boulders of resistant Lower Cretaceous sandstones have remained on the foreshore in extensive residual aprons. These aprons armour the shoreline and dissipate wave energy, performing a vital function in maintaining the stability of the Undercliff for some 5,000 to 6,000 years since early Holocene rising sea-levels initially reoccupied the slope toe. A provisional chronology of instability, characterised by major and minor falls and slides, through to the modern period has been established from archaeological evidence and radiocarbon dating of organic horizons trapped between successive ground movements.
4. Historically, much of the Undercliff- with the exception of the sector between Blackgang and St. Catherine’s Point- has remained relatively stable, but over the past seventy or so years its stability has reduced and ground movements have increased in frequency at Bonchurch, in parts of Ventnor, St Lawrence and at Niton. Although stability is related closely to groundwater conditions (critical pore water pressures at shear surfaces), it is likely that millennia of toe erosion have also reduced the support afforded by the debris accumulation of the lower slopes.
5. In recent decades there has been a definite tendency towards reactivation of specific landslide units on the lower slopes that then have had "knock-on" effects upslope such that instability has progressed to, or close to, the toe of the landward backscar developed within in situ Chalk and Greensand between St Lawrence and Niton. Increased periodicity of mass movement events since the mid twentieth century appears to directly correlate with more frequent high rainfall magnitudes and intensities.
South West Coast
1. Rising sea-levels of the mid to late Holocene re-occupied former degraded cliffs initiating renewed erosion of its soft Cretaceous sands and clays to form a rapidly retreating linear or slightly embayed cliff coastline some 15 km in length. As the coast retreated it produced a shallow nearshore shelf, or shore platform, extending seaward for some 4km which is thought to indicate the extent of mid to late Holocene coastal recession.
2. Recession has been controlled partly by the occurrence of more resistant strata forming the northwest (Chalk) and southeast (the Undercliff boulder aprons) extremities of this segment.
3. The steadily retreating coastline has eroded much of the catchment of the northward flowing Western Yar River; its higher order tributaries, possibly evident today as steeply incised chines are arguably the last remnants of this former drainage system south of its present source just meters north of Freshwater Bay.
4. Although significant volumes of material would have been released as a result of such rapid recession along a wide front, the majority of sediment yielded would have been clays and sands that were rapidly removed offshore by wave and tidal action.
5. Variations in cliff morphology and style of recession would have developed along this unit as a result of variations in ground elevation, lithology, stratigraphy and geological structure revealed as the cliffs retreated.
6. Minor headlands have developed at Hanover and Atherfield Points due to local occurrences of harder lithologic units that have formed protective foreshore reefs. However, the rates of cliff retreat are such that formerly more pronounced salients of this type have had limited longevity. Analysis of Coastal Monitoring Programme lidar and aerial photography data indicates that the minor headlands of Hanover and Atherfield Points are eroding at a faster rate the adjacent frontage.
1.2 Wave Climate
Whilst geological factors control variations in the scale, erosional resistance and detailed morphological character of the cliffed coastline, changes in coastal orientation introduce contrasts in exposure to wave energy. The west-facing coastline is open to high-energy Atlantic swell waves from the southwest that can propagate across a fetch distance in excess of 4,000km. The well-documented history of shipwrecks along this largely unprotected rugged coast is a testimony to this fact. HR Wallingford (1999), calculated, using numerical modelling of synthetic data for wave climate that the maximum wave height, for a 1 in 1 year recurrence, is close to 5m for the coastline between Freshwater Bay and the Needles. For Compton Bay, it is 4.26m. Estimations for longer recurrence intervals are also given. Variation is due to the range of different wave types and approaches. The south-facing coastline has a maximum fetch of 150km, determined by the opposing Channel coast of France, and is affected by refracted ocean swell from the west and southwest and partially refracted and diffracted waves propagated from the east and south-east. Offshore wave heights here are depth-limited by extensive submerged platforms and boulder aprons, resulting in maximum heights of 2.6m at a 1 in 5 years recurrence (Rendel, Palmer and Tritton, 1993).
By contrast, the east-facing coast is relatively protected from waves generated by dominant westerly or southwesterly winds, although subject to the residual energy of swell waves refracted by a combination of offshore seabed topography and the acute change in coastal plan at Dunnose. However, the east coast is fully exposed to a 170km fetch extending east and east-south-east; this can propagate waves up to 3.8 m height in association with infrequent easterly gale-force winds (Hydraulics Research, 1977b; 1984; 1991; HR Wallingford, 1992; Royal Haskoning, 2010). Six major storms occurred during the winter of 2013/14, with a mean maximum wave height of 3.2m measured by a buoy 1.5km offshore Sandown. Thus, no part of this coastline is immune to the effects of storm waves of exceptional energy with a high potential for effecting substantial erosion of both cliffs and beaches [Sandown Bay Wave Buoy data 2004-12].
The Coastal Monitoring Programme measures nearshore waves using a network of Datawell Directional Waverider buoys. Between 2003 and 2012, the buoy deployed at Sandown Bay in 11m CD water depth, confirmed the prevailing wave direction is from the south, and an average 10% significant wave height exceedance of 0.98m. The buoy deployed at Milford-on-Sea in 1996 in 10m CD water depth, recorded prevailing wave direction is from southwest-by-south and an average 10% significant wave height exceedance is 1.31m. The buoy deployed at Hayling Island in 10m CD water depth, confirmed the prevailing wave direction is from south-by-west, and an average 10% significant wave height exceedance of 1.26m. The buoy deployed at Bracklesham Bay in 10 m CD water depth, from 2008 to 2012, confirmed the prevailing wave direction is from southwest–by-south, and an average 10% significant wave height exceedance of 1.47m (CCO, 2012).
Freshwater and Sandown Bays were two of the locations for which wave modelling exercises were undertaken as part of the DEFRA Futurecoast Project (Halcrow, 2002). Offshore wave climates were synthesised based on 1991-2000 data from the Met Office Wave Model and then transformed inshore to prediction point in Freshwater Bay at -3.83mOD and Sandown Bay at -4.44mOD. At Freshwater, results indicated that the major approach direction for waves was between 180 degrees and 210 degrees with the highest waves typically up to 3m and a significant swell wave component (with periods greater than 8 seconds). At Sandown, results indicated that the major approach direction for waves was between 150 degrees and 180 degrees with the highest waves typically up to 2.5m and little swell wave exposure. The lesser impact of the latter would be attributable to the shelter provided by Dunnose headland so that waves arriving from west of 180 degrees would be refracted and diffracted around it. Potential sensitivities to likely climate change scenarios were then tested by examining the extent to which the total and net longshore energy for each scenario varied with respect to the present situation. Results suggested that a one to two degree variation in mean wave direction would have relatively little effect at either site. Wave energy was, however, found to be sensitive to sea-level rise with up to a 20% increase estimated for a high scenario of sea-level rise. The effect is probably due to a reduction in shoaling and wave refraction within the shallow nearshore bed as water depths increase so that slightly higher waves might approach the shoreline at rather more oblique angles. Freshwater Bay was also found to be sensitive to an increase in Atlantic storminess, with up to a 5% increase in energy likely (according to the scenario tested). Sandown, by virtue of its more sheltered position, was not considered sensitive.
1.3 Human Influences
The character of this coastline also reflects the history of defence and protection, which in turn relates to land and resource development over the last 150 years. The west coast, with the exception of Freshwater Bay is, and always has been, largely devoid of protection structures. Thus cliff development has been unrestrained, and the beaches here are in adjustment to natural sediment supply from available sources e.g. (Photo 1). The western and central sectors of the south, or 'Undercliff', coast have also experienced little modification by human activity (Photo 2). This is equally the case for the southeastern coastline between Dunnose and Horse Ledge (Photo 3) and the Chalk promontory of Culver Cliff (Photo 4). By contrast, the remaining coastal frontage, especially between Ventnor (Photo 5) and Monk's Bay (Photo 6) and the east coast between Shanklin Chine (Photo 7) and Yaverland (Photo 8), is fully protected by a variety of structures. These include sea walls (Photo 9; Photo 10 and Photo 11), revetments and several groyne fields (Photo 7 and Photo 12) that have been subject to both renewal and extension for more than a century. The groyne system between Shanklin and Sandown (Photo 12 and Photo 13) has succeeded in retaining substantial quantities of sand, transported from south to north by the net direction of littoral drift. Along the undefended natural coastline, beach character exhibits rapid spatial changes, from sand through various grades of gravel to boulders.
There would appear to be a direct relationship between sediment released from sites of active cliff recession and landslide reactivation of the coastal slope and the character of beaches short distances downdrift. There are, however, uncertainties over the sources of supply to the 'pocket' beaches located in isolated bays and coves along the southern and south-eastern coastlines, e.g. (Photo 14) and a general lack of reliable information on offshore to nearshore and inshore sediment supply.
Much of the literature on the details of coastal processes is in the form of unpublished consultants' reports. There remain several sectors of the shoreline for which process information is scarce, largely because lack of commercial, residential and infrastructure development has not hitherto required formal shoreline management. (Halcrow, 1997; Brampton, et al., 1998). The policies of the Isle of Wight Council are in favour of maintaining the natural environmental attributes of this coastline (McInnes, et al., 1998).
2. Sediment Inputs
2.1 Marine Input
F1 Onshore Feed
The offshore to onshore supply of sediment by wave-induced or tidal currents may account for a proportion of beach stores at certain locations. However, knowledge of nearshore sediments and possible pathways of transfer to littoral transport is very limited and is largely a matter of conjecture (Brampton, et al., 1998). It is known that parts of the shoreface between The Needles and St Catherine's Point are current-swept bedrock surfaces (Posford Duvivier and British Geological Survey, 1999), thus implying limited supply potential. Tidal currents achieve relatively high velocities of 1.5 to 2.1ms-¹, and flow sub-parallel to the coastline. They may effect scour around large boulder accumulations and gravel patches.
In 2011 the Coastal Monitoring Programme completed an high resolution, 100% coverage swath bathymetry survey of the nearshore zone of the northern and southern coasts of the Isle of Wight, extending 1km offshore from the MLW. This southern survey area abutted a survey further offshore commissioned by Natural England, which completed the entire coverage of the extent of the marine Special Area of Conservation.
The high wave energy nearshore zone of the southwest Isle of Wight is characterised by extensive areas of exposed bedrock, extending from the toe of the cliffs and inter-tidal rock platforms, notably offshore of Freshwater Bay and Sandown Bay. These rock outcrops are interspersed with constrained areas of sand and sandy gravels, which are largely of insufficient thickness to mask the underlying bedrock.
Between the Needles and Afton Down the southerly extension of the cliff toe and rock platform outcrops offshore approximately 500m, maintaining the west-east sinuous alignment of the cliffs. Further offshore sediments are sufficiently thick to cover the underlying bedrock. This area of sediment extends from the Needles eastwards onto Compton Beach. The rock platform is then exposed towards the southern portion of Compton Bay. The nearshore and sub-tidal zone of the exposed south west facing coastline between Compton Bay and Atherfield Point is dominated by exposure of the rock platform, which extends offshore from the base of the cliffs. The bathymetry data provides no conclusive evidence of indicative onshore transport, as shown in 2004 maps, and arrows offshore of Chilton Chine and in Brighstone Bay have been removed. Where accumulations of sand and mixed sediment occur, such as south of Freshwater Bay and between Atherfield and Walpen Chines, these are largely constrained in sparse pockets within the sub-tidal rock platform. There is a lack of bedforms or evidence to support onshore pathways of sand and gravels from offshore sources, postulated by Posford Duvivier 1989a, 1990a and b; Barrett, 1985; Kay, 1969; Rendel, Palmer and Tritton, 1993; Halcrow, 1997; Brampton, et al., 1998).
There is higher occurrence of nearshore sediment extent and thickness between Atherfield Point and Rocken Point, sufficient to cover the rock platform in places or cover it in a thin veneer of sediment. From Rocken Point southwards and around St Catherine’s Point the rock platforms and exposed outcrops are more evident. The nearshore seabed between St Catherine’s Point to Monks Bay is a complex mix of rock platform and outcrops close to shore and rock covered in varying thicknesses of sediment further offshore. Small constrained patches of sand and coarse sediment are evident in numerous pocket beaches between St Catherine’s Point and Luccombe Bay including Reeth Bay, Ventnor Bay and Monks Bay. Offshore of Ventnor Bay there is an area of large sedimentary bedforms, the majority of which are slightly asymmetrical indicating west to east sediment transport flow, although further offshore are smaller-scale slightly asymmetrical bedforms that indicate east to west transport.
2.2 Fluvial Input
FL1
It is possible that sediment is discharged as traction or bedload by the several chines that interrupt the continuity of the cliffline of the western and southeastern shorelines. See May (2003) for a concise review of the several hypotheses of chine formation and development. Flint (1982) suggests that most of the debris transported by the chine streams of the west coast are in the fine sand and silt fractions, though her conclusion is not based on sampling. The bedload of some of the more deeply incised valleys is coarse, derives from large ironstone doggers exposed in valley-side slopes, and indicates the role of stream power. Whale and Brook Chines have boulder chokes at their mouths, indicating occasional high discharge and bedload-transport competence. However, the petrographic character of this material is an unlikely source for durable beach gravel. Nonetheless, local beaches may retain a small proportion of the sediment discharged by the chine streams. Human interference with discharge: load ratios may have affected this contribution in the last century or more, as Flint (1982) notes that Shepard's Chine, north of Atherfield Point, has cut down at least 15m since 1820 as a result of local diversion of drainage. Works of art, some dating back to the eighteenth century, portraying several of the island chines reveal striking morphological changes resulting from incision due in part at least to base-level change induced by cliff retreat (McInnes, 2008). Rendel Geotechnics and University of Portsmouth (1996) calculate that the West coast chines collectively transport 803 tonnes per year of suspended load and 259 tonnes per year of bedload. Of these quantities, only 82 tonnes per year, is delivered to the coastal transport system; the remainder is diverted to various forms of channel storage, both naturally and artificially induced.
2.3 Coast Erosion
Introduction
Analysis of Coastal Monitoring Programme 2006 to 2012 lidar, 2003 and 2012 aerial photography and topographic baseline survey data, combined with other datasets, academic research and historical studies has enabled sediment budgets, transport rates and directions to be identified or verified.
The entire length of this coastline is subject to active shoreface and inter-tidal marine erosion and cliff development, with the exception of sites such as Shanklin and Ventnor where sea walls and promenades have removed the former cliffline from the direct influence of wave-induced attack. Vertical shoreface erosion rates of between 3 and 10 mm per year for the Undercliff, and 1.2 to 4 mm per year along the coastline north of Shanklin Chine are estimated by Posford Duvivier and British Geological Survey (1999). Cliff morphology is spatially varied, and reflects differing combinations of factors, the main ones being:
- Rock lithology and internal structures, principally intersecting joint sets, failure planes and faults;
- Stratigraphic succession of contrasting rock lithologies, permeabilities and porosities;
- Local dips consequent upon regional geological and tectonic structure;
- Wave climate, including localised effects on wave refraction created by nearshore and offshore water depths and seabed topography;
- Dynamics of adjacent beaches and basal debris stores;
- Weathering and mass movement (sub-aerial) processes, which may partly depend on microclimate and vegetation colonisation (including human modifications of both), as well as rock/sediment characteristics;
- Hydrogeological conditions, which are largely a function of (a) (b) and (c), but which may also reflect drainage towards the coastline from the immediate hinterland feeding the rear of active, reactivated and relict landslips;
- Holocene climatic history.
The above listing does not imply an order of importance of these factors. There are certain sectors, notably along the Undercliff coast (Photo 2), where it can be argued that the geomorphological history of cliffline instability is the most important and direct input into contemporary morphological character. Hutchinson (1987, 1991); Hutchinson and Bromhead (2002) and Moore, et al., (2007) have revealed striking spatial variation in the magnitude and frequency of landslide and rockfall events in recent centuries. These events have determined the dimensions of basal debris aprons and, in turn, some of the salient features of the plan form of the coastline. Moderate to high rates of cliffline retreat are characteristic of much of this coastline, evidenced by well-developed shore platforms between Shippards Chine and Brook Bay on the west coast; Horse Ledge and Luccombe Bay in the south-east; and offshore Culver Cliff. The stacks offshore the Chalk cliffs at Freshwater Bay, as well as the Needles (Photo 15), provide further dramatic proof (May, 2003). For all of these areas hydrographic chart data indicate that shallow water (less than 20m depth) extends as much as 4km offshore. Several authors (e.g. Steers, 1964; Devoy, 1987) have cited the truncated form of the River (West) Yar, particularly the proximity of its southern catchment boundary to the beach at Freshwater Bay, as evidence of the rapid recession of the west coast throughout the Holocene period of sea-level rise. It is a reasonable deduction that the West Yar drained a substantial catchment basin now lost to coastal erosion. The coastal chines, or valleys, incised into the cliffline of the west coast have been interpreted by earlier authors (e.g. Jukes-Browne, 1874; Osborne White, 1921) as the remnants of its headwater tributaries. The plan form, and alluvial deposits, of the stream discharging via Brook Chine provides evidence for this hypothesis, although Flint (1982) has proposed that the chines might be a product of dynamic balance between basin area, stream power and the rate of cliff retreat since the relative stabilisation of sea level at the end of the mid- Holocene transgression circa 5,300 years BP. Despite their different orientations, the same arguments must also apply to Luccombe and Shanklin Chines on the south-east coast.
Extrapolation of measurements of coastal recession for the past 150 years (e.g. Posford Duvivier, 1989a, 1999; Halcrow, 1997; Tomalin, 1977; May, 2003) supports the conclusion that there has been up to 6km of retreat of the western coast since the start of Holocene sea level recovery between 12,000 and 11,000 years BP. This estimate can be applied with most confidence to those sectors where there are outcrops of comparatively weak, erodible sandstones, clays, marls and interbedded limestones. Along the Undercliff, erosion of the landslide toes has released residual aprons of basal boulder-sized material and created a series of small bays protected by the aprons, so that there has been some localised enhancement of cliff resistance to wave energy. This may also be the case for (i) the coastline from Luccombe south to Monk's Bay, where cliff erosion releases large inter-joint blocks that litter the inter-tidal zone (Photo 16); and (ii) the Chalk coast in the extreme eastern and western sectors. The cumulative effect of successive major and minor landslides in parts of the Undercliff has been to temporarily advance the position of parts of this coastline, but the distance of erosional retreat over the past 5 or 6 millennia is probably in the order of 2 to 3kms. Since the mid Holocene wave erosion has been concentrated within a relatively narrow height range, which has promoted the expansion of shoreline platforms at suitable locations. Their effect may have been to dissipate the potential assailing force of breaking waves and thus slowly decelerate coastline recession rates. This, and other, factors have not been quantified for this coastline, but they serve to qualify calculations of land loss based solely on contemporary rates of shoreline retreat (Halcrow, 1997).
Whatever the longer-term variations in rates of erosion may have been, substantial quantities of sediment have been released from cliff erosion. This must be a primary source of the sediment, mobilised principally by waves, which has contributed to beach stores. An additional input, from offshore, is also a possibility (Brampton, et al., 1998), however, a lack of detailed study information regarding cliff sediment composition and proportions of sediment grain sizes means that it has not been possible to quantify volumes of cliff input yielding shingle or sand grade beach material.
The Undercliff Coast: Chale to Dunnose
The entire Undercliff coast is fronted by some twelve compound slides of landslip debris, mostly comprising massive multiple blocks of Chalk and Upper Greensand with slip planes seated in the underlying, gently seawards dipping Gault Clay and Lower Greensand (Photo 2). The geological and geotechnical properties and both past and present morphological character of the coastline have been described in detail elsewhere (Hutchinson, 1965; Hutchinson and Bromhead, 2002; Hutchinson, Chandler and Bromhead, 1981; Hutchinson, 1983, 1987; Lee, Moore and McInnes, 1998; Geomorphological Services Ltd, 1991; Hutchinson, Brunsden and Lee, 1991, 1994a; McInnes and Jakeways, 2000; Tomalin, 2000; McInnes, 2007; Rendel Geotechnics, 1994, 1995a, 1995b, 1996, 1997; Rust and Gardener, 2002; Posford Duvivier, 1991, 1993c, 1994a; Moore, et al., 2007, 2010; Gillarduzzi, et al., 2007; Palmer, et al., 2007 a and b), as part of a comprehensive series of studies of the magnitude and frequency and causal relations between factors inducing instability; the socio-economic impacts on property, infrastructure, and residents and future management options and strategies of this extensive landslide system (McInnes, 2000; 2007). These several sources provide detailed geomorphological and geo technical maps; data logs and documented details of monitoring of ground movements at both the coastline and inland. The principal processes that have contributed, in varying relative proportions, to spatial variation in the scale and types of slope instability and coastal erosion are "run out" and rotational slips (slope and base failures); translational slides; rockfalls; mudflows and seepage erosion. Erosional cycles of unloading, steepening and debris accumulation, from 100 to 150 years duration to periods of several centuries and millennia, affect both basal and higher slopes (Rendel Geotechnics, 1995; Ibsen and Brunden, 1996; Ibsen, 2000; Bray and Hooke, 1997; Tomalin, 2000; Hutchinson and Bromhead, 2002). In general terms, compound translational slides dominate the seaward sector of the Undercliff, with multiple rotational slides located inland. The contribution of direct wave energy impact at the cliff base and offshore seabed relief to the initiation and acceleration of slope failures has yet to be precisely quantified, but these factors undoubtedly account for some of the contrasts in the longer-term stability of different units of the Undercliff coast as a whole. It is also becoming clearer, with improved monitoring data and extension of the local climatic record, that periods of prolonged (or events of intensive) rainfall, especially when effective winter precipitation exceeds 400mm over three to four months, are also a significant trigger of ground movement and slope failure to slopes over-steepened by coastal erosion. Analysis of historical and recent hydro-climatic data back to 1839 (Ibsen, 2000; Moore, et al., 2007, 2010; Gilladuzzi, et al., 2007) has demonstrated a direct relationship between antecedent rainfall, the elevation of groundwater levels, critical pore water pressures and episodes of ground instability. The potential for landslide reactivation is also likely to be enhanced by accelerated seacliff recession due to sea-level rise. Debris storage is influenced by the relatively narrow shoreface, with the 10m isobath approaching to within 200m of the coastline. St Catherine's Deep, exceeding 60m in depth, is less than 2km offshore, and is bounded by landslip derived boulder-covered slopes from which fine material may have been removed by wave scour and tidal currents. This 35km long depression is controlled by geological structure, and may be eroded into a Gault Clay Outcrop. It may be, in part, an original palaeo valley partly deepened by tidal scour (Hutchinson and Bromhead, 2002).
The "classic" component of the Undercliff coast, between St Catherine's Point and western Ventnor, is developed in the full succession of Lower Greensand, Gault Clay, Upper Greensand and Chalk. The regional dip is 1.5-2.5° to the south and south-south-east. A strongly defined 60-80m free face in the Chalk and Upper Greensand provides a backscar to the landslip complex, which is made up of a series of terrace-like features. These are irregular in width and relative height (70-110m) although most are elongate in plan. Each terrace is a large rock mass that has moved seawards across a deep-seated semi-rotational failure plane. Multiple rotational slides dominate the upper part of the Undercliff, with several back-tilted blocks of Upper Greensand in positions of longer term but temporary stability (Hutchinson and Bromhead, 2002; McInnes, 2007). A zone of similar width occupies the lower Undercliff, consisting of a sequence of compound slides; some of these give rise to ridge-like forms defining small grabens (areas of relative subsidence). The failure planes (basal shear surfaces) are located within both the Gault Clay and clay-rich horizons in the underlying Sandrock, and have different configurations for various parts of the Undercliff (Geomorphological Services Ltd, 1991; Rendel Geotechnics, 1995a, b). Geotechnical explanation of failure principals involves the contrasting hydrogeology of the Upper Greensand-Gault Clay-Carstone-Sandrock sequence (Hutchinson, 1991; McIntyre and McInnes, 1991; Hutchinson and Bromhead, 2002; McInnes, 2007; Moore, et al., 2007; Moore, et al., 2010). Monitoring of borehole inclinometers since 1992 has indicated deep-seated movement of 2 to 3mm. per month during dry periods, increasing to over 30mm immediately following events of heavy winter precipitation (Rust and Gardener, 2002; Gillarduzzi, et al., 2007).
A relatively continuous steep scarp slope, up to 20m in height, defines the Undercliff terrace west of Ventnor. It is underlain by Gault Clay and is the site of occasional mudslides. Degraded mudslides, also in Gault Clay, are located immediately above cliffs cut into the Lower Greensand, e.g. at Binnel Point. The Gault Clay scarp is regarded as a major influence on the mechanisms of slope failure within the Undercliff as a whole. Its intermittent unloading, and retreat, has triggered rotational failures that, together with slope movements on the lower Undercliff, have contributed large aprons of landslide debris. During pre-Holocene times, these debris stores would have provided basal loading, and in the early stages of Holocene sea-level rise, natural protection against erosion. Progressive removal of this apron of material resulted in slope instability (Photo 24 and Photo 25), either creating new failure planes or re-activating old ones. Curved backscars in the lower landslide units appear to have been a direct response, thus isolating a series of broad, roughly triangular, spurs. These have also failed as a result of unloading on either side (Rendel Geotechnics, 1995a). An inverse relationship may exist between the degree of preservation of debris aprons and coastal slope sensitivity to shoreline conditions.
In general terms, the types and scales of past and current slope instability suggest a two tier cascading system involving a combination of different failure mechanisms at different elevations (Moore, et al., 2007, 2010). The lower elevation ("Zone 1") failures are principally translational, and are controlled by clay horizons in the Sandrock (Lower Greensand). Above these, "Zone II" failures are of the multiple rotational type, related to slip surfaces in the Gault Clay. Characteristic morphology in Zone I is of one or more sub-parallel linear ridges, whilst relatively narrow terrace-like flats separated by well-defined scarps are diagnostic of Zone II. In places, such as the vicinity of St Lawrence, debris aprons partly overlie ridges, both of which consist of landslide debris in slow downslope transit. At the extreme eastern end of the complex, mudslides and recent landslides replace the Zone I landform assemblage as a consequence of recent, and continuing, active slope movements.
The evolution of this spatially variable geomorphic pattern is related to the inter-relations between progressive removal of formerly larger and more extensive landslide debris stores as sea-level rose and the coastline retreated. Deep-seated failure planes developed in Zone I in response, creating several isolated triangular shaped spurs. Their defining slopes subsequently failed, due to erosional unloading, thus initiating the Zone II backscars. For all parts of the unprotected Undercliff frontage, marine erosion is therefore a critical control of both slope instability and sediment supply to the littoral transport system.
This frontage is dominated by high, near vertical up to 140m high Chalk cliffs with a few small coves and slight indentations that accumulate small fillets of coarse angular gravel (Photo 15 and Photo 17). Of these, Scratchell's Bay is the largest, but is inaccessible. Basal shore platforms are poorly developed. The Needles are stacks resulting from marine erosion of tensional joints. Two additional pinnacle stacks were destroyed in 1706 and 1772. Barrett (1985) has made the general observation that cliff recession rates appear to be least where there is little, or no, beach accumulation. This would appear to imply that wave-assisted clast abrasion rather than direct cliff toe erosion along this exposed coastline is a significant factor. Beach material is presumed to derive directly from the release of flint nodules from the steeply-dipping bedding planes of the Upper Chalk at a rate of 1500 m³ per year (Posford Duvivier, 1997, 1999) from a total yield of 15,000m³ per year of chalk debris. Recession of this cliff line is relatively slow, with intermittent rockfalls evidenced by boulder accumulations that obscure much of the cliff toe. A rate of shoreline recession of 0.14m per year is suggested by Posford Duvivier (1991) and 0.15m per year was calculated by Halcrow, (1997) covering the period 1866-1995. The shoreface is relatively steep, with the 10m isobath between 200 and 300m from the shoreface. This limits capacity for debris storage. Analysis of the Coastal Monitoring Programme indicates some cliff-derived sediment is retained on the shore platform, with fine grained or smaller blocks of Chalk are removed offshore in suspension. The lack of detailed study information regarding cliff sediment composition and proportions of sediment grain sizes means that it has not been possible to quantify volumes of cliff input yielding shingle/flint beach material.
Field and documentary evidence suggest that the prevailing rate of retreat of the Chalk cliffs is slow due to the high degree of protection afforded by the beach - approximately 0.08m per year (May and Heaps, 1985; Barton and McInnes, 1988). Very localised and episodic retreat of parts of the cliff-top free face occurs at rates of up to 0.3m per year (McInnes, 1994). Prior to the provision of coastal defences within the bay, recession occurred at a mean rate of around 0.5m per year for the period 1866-1909 (Halcrow, 1997). Occasional rockfalls yield a small quantity of clastic material, but in the western part of the bay the cliffs support an overburden of loosely consolidated chalk and flint fragments (Coombe Rock) that is subject to mass wasting and gullying. Posford Duvivier (1999) calculate an erosion loss of approximately 2000m³ˉper year of mixed sediment sizes, of which less than 100m³ˉper year is flint gravel that is retained on the local beach. Analysis of the Coastal Monitoring Programme indicates the erosion of the cliffs within Freshwater Bay supplies less than 1,000m³ˉper year of beach grade material (the 2004 arrow indicated no quantitative data). Cliff morphology suggests that mechanical failure due to sub-aerial denudation, toe erosion, and block fall are in phase (Geodata Institute, 1989). The presence of the sea stacks (including Stag rock, and the former Arch rock, which collapsed in 1992) is proof of longer term erosion of the more exposed cliffline; a third stack, Mermaid rock, is not in situ but added as a result of a large block fall in 1971.
Between Freshwater Bay and Compton Down, Chalk strata dipping at 45 to 80 degrees form cliffs in excess of 75m height. Material at the cliff foot as well as cliff top tensional fissures suggests that gravity-assisted falls and block failure are dominant here (Barton, 1990). May (1966) calculated a rate of shoreline recession of 0.01m per year, which is significantly less than for Chalk cliffs of similar exposure and dimensions at other south coast locations. Posford Duvivier (1981, 1989a, 1997, 1999) propose an average long-term rate of 0.15-0.6m per year, yielding some 15,000m³ˉper year of Chalk and 500 m³ˉper year of flint gravel. Analysis of Coastal Monitoring Programme indicates less than 1,000m³ˉper year of beach grade material derives from the erosion of the cliffs in the northern section of Compton Bay, a reduction in volume compared to the 2004 proposed rate of 3-10,000m³ˉper year. Halcrow (1997) calculated a long-term recession of around 0.1m per year from 1886 to 1975, but with an increase to 0.42m per year for 1975-1995 attributable to failures in superficial deposits at the cliff top. Recession of the cliff top at Afton Down has posed a particular threat to the A3055 since 1981. The cliff slope hereabouts is partly defined by four sets of east-to-west trending joints, but is mantled by slip debris derived from Chalky Head materials above. Detailed geotechnical surveys, the installation of a pile retaining structure and monitoring using tilt meters and extensionometers, have been undertaken since 2007 because of the threat to public safety (McInnes and Jakeways, 2000; Clark and Fort, 2009; Wood, et al., 2004). Barton and McInnes (1988) calculate a recession for the cliffs as a whole of 0.05m per year with up to 18m of retreat of the cliff top between 1935 and 1980 (McInnes, 1983, 1994). They suggest that the cracks or "vents" that have appeared are indicative of incipient toppling failure, ultimately related to toe erosion. Movement has been especially noticeable during winter months, related to either the swelling of superficial drift, "ponding" of perched groundwater aquicludes or pressure-release due to previous failures (Barton, 1987; Geodata Unit, 1989).
However, the overall erosion rate of about 0.10m per year along this sector of the coastline is comparatively slow, and may be partially explained by the exceptional width of the offshore zone. The 10m depth contour in the north of Compton Bay indicates a 1300m wide offshore platform so that incident wave energy is strongly dissipated. However, the Chalk yields very little sediment suitable for beach building, so that protection against breaking waves is slight.
To the south of Compton Down, the lithology of the outcropping Greensand and Wealden rocks includes clays, shales, marls and sandstones. The offshore platform is wide, but modified refracted wave erosion is partly offset by the highly erodible nature of the cliff materials. Overland flow, gullying, slump failures and shallow sliding contribute to the constantly changing morphology of cliff faces, and can be important alongside basal marine erosion in regulating retreat rates. Hanover Point owes its existence to an outcrop of comparatively more resistant bedrock; this factor, and the width of the intertidal platform at this point, has suppressed erosion. Analysis of the Coastal Monitoring Programme 2001, 2008 and 2013 aerial photography data indicates significant cliff erosion, with an annual recession rate of 0.8m at the Hanover Point headland. Posford Duvivier (1981, 1989a, 1999; McInnes, 1994) report a contemporary rate of retreat of 0.2-0.5m per year, with 24m of recession at Shippard's Chine, 1877-1955 (Isle of Wight County Surveyor, 1957). The volume of beach grade material derived from the erosion of the cliffs is estimated at 3-10,000m³ˉper year, a reduction in volume compared to the 2004 proposed rate of more than 20,000m³ˉper year. An annual recession rate of 0.6m at Shippard’s Chine has been determined from Programme data. For the complete section between Compton Chine and Hanover Point, Halcrow (1997) calculated recession averaging some 0.48m per year for the period 1866-1995, although this value includes considerable local variation. A rate of 0.23m per year at Compton Chine is suggested by Barton (1987). McInnes (1983) has proposed a figure of 4m per year in "recent" years at one particular site where there has been particularly active mass wasting and possible seepage erosion. Barton (1987) has calculated 23m of land loss, 1861-1981 for the narrow Gault Clay outcrop. Rust and Gardener (2002) calculated a mean rate of 0.9m per year for the sector between Hanover Point and Compton Bay, 1862-1976. There is a well-defined shoreline platform between Shippard's Chine and south of Hanover Point cut across both weak and more resistant bedrocks, providing clear proof of the long-term efficacy of linear erosion at an estimated rate of 0.55m per year (Posford Duvivier, 1999). Between Afton Down and Shippard's Chine Posford Duvivier (1997) proposed a total sediment yield of 35,000m³ˉper year dominated by sand, silt and clay. Between Shippard's and Grange Chines, an erosion yield of 45,000m³ˉper year of fine sands and clay and 5,000m³ˉper year of gravel-sized sediment was estimated.
With the exception of the weakly arcuate plan of the inset of Brook Bay, this comparatively straight coastline is dominated by cliffs of 20-30m in height (Photo 18), rising to 53m at Barnes High (Photo 19 and Photo 20) and is interrupted by several chines. Geological composition overall is of weak materials, similar to the sector to the northwest but with significantly higher rates of localised cliff top recession due to several semi-rotational slides, slumps, block failures and mudflows. Localised relatively more resistant rock outcrops coincide with cliff face steepening. Over the past 20 years, the highest rates of retreat have been along the 1km frontage south of Brookgreen. A shore platform extends to just south of Chilton Chine (Photo 1), but is frequently concealed by both beach sediments and rafts of seaweed. A detailed study of shoreline movement in the immediate vicinity of Grange Chine gave a figure of 130m of recession, 1794 to 1977 (1.6 m per year) (Tomalin, 1977) with possibly up to 700m of retreat "since Roman times". McInnes (1983) derives a figure of 1m per year, 1935-1982 for the low sandstone and clay cliffs in Brook Bay. Colenutt (1938) wrote that "extensive erosion has occurred north of Brook", presumably referring to immediately preceding years; he infers that this was due to storm waves gaining access to the cliff foot, which had previously been protected by a wide, stable sandy beach. May (1966) calculated 49m of retreat within Brightstone Bay between 1870 and 1963, giving a mean of 0.52mper year, but in excess of 1.0m per year where rock resistance is reduced (May, 2003). The County Surveyor (1957) derived a similar figure, 0.55m per year, for Brook Bay, whilst Barrett (1985) proposed an average recession rate of 0.4m per year since 1800 and Posford Duvivier (1997) suggest 0.48m per year. Rust and Gardener (2002) derive a rate of 0.3m per year for the period 1860 to 1976 based on their analysis of successive OS map revisions. For the sector between Hanover Point and Sudmoor Point, Halcrow (1997) calculated recession averaging some 0.40m per year for the period 1866-1995. They also calculated recession averaging 0.51m per year for the sector between Sudmoor Point and Grange Chine and 0.47m per year between Grange Chine and Sheppard’s Chine. This range of values may reflect genuine spatial variations in wave energy induced by platform relief (e.g. Brook Ledges), cliff height and rock erodibility, but may also be due to temporal variations in the magnitude of coastal storms. These rates translate into a supply of 105,000m³ˉper year of mostly fine sediment between Shippard's and Shepard's Chines (Posford Duvivier, 1997); of this, less than 5,000 m³ˉper year is coarse material. Coastal Monitoring Programme data indicates areas experiencing landslide and cliff top recession, with majority of eroded material retained on the coastal slopes and cliff faces, with a foreshore input of 3-10,000m³ˉper year, a reduction in volume compared to the 2004 proposed rate of more than 20,000m³ˉper year.
The position of the 10m isobath along this straight, actively eroding cliffed coastline (interrupted by two major chines) approaches to within 500m of mean low spring tides, and the effect of Atlantic swell waves is greater here than on adjacent parts of the west Wight shoreline. Atherfield Point, and its offshore ledges, results less from an outcrop of a more resistant unit within the Lower Greensand than the width of the adjacent intertidal shore platform (Photo 21 and Photo 22). The cliffs are formed primarily within the Ferruginous Sands and Sandrock and maintain near-vertical profiles with occasional narrow ledges (Photo 21, Photo 22 and Photo 23). Long term recession rates of 0.6m per year (Posford Duvivier, 1997) and 0.66m per year for 1866-1995 (Halcrow, 1997) for this unit, with spatial variability due to subtle changes in rock resistance and exposure to wave energy, are characteristic. The deeply incised profile of Whale Chine can be explained in part as a result of rapid cliff retreat (i.e. base level lowering due to shoreline recession) and in part to an earlier event of stream capture. An erosion yield of 45,000m³ˉper year between Shepard's and Whale Chines is derived from the above recession rate (Posford Duvivier, 1997). Although cliff erosion has continued, it has not been possible to quantify the volume of beach grade material derived from the erosion of the cliffs, although the 2004 proposed rate was estimated at more than 20,000m³ˉper year.
Cliff height increases from 70m at Chale (Photo 23) to over 180m at Gore Cliff (Photo 26), where the eastwards dip of the rock succession causes a west to east change in the lithological units outcropping at the cliff base. The Ferruginous Sandstone between Walpen Chine and north of Rocken End, but particularly in the vicinity of Chale, contains bands of interbedded clays that promote shearing and seepage erosion of the overlying sandstones above the wave eroded cliff base. Associated aquicludes promote seepage erosion and are responsible for the bench-like form of the cliff profile. Study of Ordnance Survey map editions between 1861 to 1960 give an overall recession rate of approximately 0.5m per year. Kay (1969) noted that the true Undercliff, defined by slip and flow debris beneath an upper scarp and bench, was widest between Cliff Terrace and Blackgang (Photo 27). He reports 25.3m of upper scarp retreat between 10 December 1965 and 19 February 1966 and comments that the building of the properties of Cliff Terrace in the 1860s might suggest a period previous to that time of apparent stability. Hutchinson (1987) reports a significant acceleration of coastline retreat at this site over the past 130 years, from 0.16m per year between 1861 and 1907, to 0.57m per year in the period 1907 to 1980, giving a mean rate of 0.43m per year. Posford Duvivier (1997) suggest a spatially averaged figure of 1.2m per year, (0.6-2.0) giving the substantial yield of 440,000m³ˉper year of sand and clay for the sector between Whale Chine and Rocken End where active cliff height is at a maximum for this entire coastline. Rendel Geotechnics (1995b) calculate a rate of cliff retreat of 1.57m per year, 1980-1994. Coastal Monitoring Programme data indicates areas experiencing landslide and cliff top recession, with some eroded material retained on the coastal slopes and cliff faces, with a small proportion transported to the foreshore, but fine grained sediments not retained. It has not been possible to quantify the volume of beach grade material derived from the erosion of the cliffs, although the 2004 proposed rate was estimated at more than 20,000m³ˉper year. Cliff top retreat throughout the 1980s and 1990s has been greater than that of marine erosion at the toe, with debris supplied by several failure reactivation events (compound slides and rotational slips) temporarily stored in the Undercliff. (Note that the coastal backscar developed in Lower Greensand materials has retreated relatively slowly, whereas the previously inactive landslides in front have reactivated towards the backscar toe). If this continues then renewed first time failures of the backscar can be anticipated in future (Halcrow, 2002). Material is removed by mudslides that are frequently released onto the beach at the cliff base, which has an ill-defined form compared to the steep cliffs to the immediate north.
The main morphological components are:
I. The Walpen (Chale) Undercliff, developed in the Atherfield Clay, Sandrock and Ferruginous Sandstone lithological units. A sequence of near-horizontal terraces and free face segments coincide with clays and sandstones, respectively. Cliff recession takes place through spatially and temporally variable combinations of falls, mudslides and erosion by groundwater seepage. Halcrow (1997) calculated long-term recession averaging some 0.62m per year for the period 1866-1995, although recession has increased in recent decades with typical rates of 0.8-0.9m per year. Historical rates of cliff top recession have exceeded those at the cliff base, but the latter provides the ultimate control. Recession of the cliff base has significantly narrowed the lower Undercliff benches; mudflows are frequently released onto the beach, but are rapidly dispersed.
II. The Blackgang to St Catherine’s Undercliff (Photo 26 and Photo 27), comprising a linked series of landslides activated by semi-rotational failure planes. Broadly, there are multiple rotational failure units in the Gault clay, succeeded downslope by mudslides developed across a mid-slope scarp and compound block failures developed on clay-rich horizons in the basal Sandrock. Steep cliffs separate each of these units (Rendel Geotechnics, 1995a), which are functionally inter-related. The unit as a whole has a history of large scale and frequent landslide reactivation, related to complex feedbacks between rates of basal cliff erosion and the unloading of debris stores as well as backscar rockfalls (Bromhead, et al., 1991; Moore, et al., 1998). Groundwater also plays a significant role, with pore water pressures at and above critical lithological boundaries probably determining the magnitude of the larger failure events (Bray, 1994). Their precise timing may be linked to exceptionally high levels of antecedent rainfall, a relationship that was apparent for the 1978 and 1994 and 2001 landslips at Blackgang (Ibsen, 2000). The 1994 event involved deep-seated rotational movements shunting a large volume of ancient landslide and rockfall debris over the Gault clay scarp. Simultaneously, several mudslides transported this material onto the basal cliff slope and beach. It illustrated clearly the integration and inter-dependence of the main morphological units, which have been analysed in terms of an evolutionary model by Rendel Geotechnics (1995a). This argued that retreat of the sea cliff reduces the lower tier block failure unit, which in turn triggers upslope landsliding through reactivation of ancient failure planes; the Gault Clay scarp is unloaded by the landward migration of compound block failures, thus promoting mud sliding. The critical control is exerted by marine erosion of the cliff foot, together with removal or reduction of the debris stores mantling the foreshore. Numerous small-scale failure events in the basal section of this system will promote episodic larger-scale shearing failures in the upper, landward zone. The precise relationship between the rate of recession of the Gault Clay scarp and upslope failure events is, however, a complex one, involving time lags that are not yet fully understood.
Despite the intermittent delivery of large quantities of landslide debris to the foreshore, some of it of boulder size, beach formation appears inhibited. The cliff toe therefore remains exposed to marine erosion.
Rates of cliffline recession have been calculated by several researchers, largely with reference to the retreat of the marine cliffs. The doubtful survey accuracy of cliff features on this coastline on successive Ordnance Survey maps renders this data potentially unreliable although it is to be presumed that the cliff top has been mapped accurately. Cliff retreat varies within and between each of the major morphological subdivisions; Rendel Geotechnics (1995a) quote a rate of approximately 2.5m per year, 1980-1984 for the Gault Clay scarp, whilst Moore, et al., (1998) report an annual average rate of 0.73m, 1861-1994 for the cliff base. This is probably representative of a phase of accelerated movement, as earlier measurements covering the period 1861-1980 (Hutchinson, et al., 1981) indicate a lower long-term mean of 0.41m per year. Halcrow (1997) calculated recession averaging 0.14m per year for the period 1866-1909, although recession increased thereafter with typical rates of 2.0m per year. It was noted that the recession process was linked to major reactivations of ancient landslides in 1928, 1935, 1952, 1978 and 1994. The episodic nature of recession at this site means that as much as 50m of retreat is possible in a single event (Bray, 1994). Minor ground movements involving tension cracks and pressure ridges extend inland. Rock falls from the backscar of Gore Cliff, as occurred in 1928, 1994 and 2001, critically load the Undercliff and thus activate mudslides. These in their turn erode fresh gullies. Relief of passive support initiates renewed rock falls (Bromhead, et al., 1991; Hutchinson and Bromhead, 2002; Barton, 2007). Mean recession rates therefore need to be considered in combination with the maximum recession likely in a single event. Sea-level rise over the next 50 years and beyond may increase access of wave erosion to the cliff base and thus sustain higher rates of recession. If so, the evolutionary model of cliff behaviour points to increased frequency of landslide events, although predictions of their locations, magnitudes and timings are not yet possible. It should be noted that this active unit provides the most appropriate analogue for the behaviour of the Undercliff should its toe protection and debris aprons be removed.
A landslide bench descends from approximately 55m OD at Walpen Chine to beach level at Blackgang Chine, with some subsidiary minor benches to the immediate south east (Photo 23, Photo 26 and Photo 27). These were apparently better defined, and by implication were more stable, throughout the nineteenth century. Over a small area immediately southeast of Blackgang Chine, the cliffs are in a relatively stable condition; Hutchinson, et al., (1981) suggests that seepage erosion is suppressed here because sub-surface drainage is directed towards the Chine. Elsewhere, groundwater-fed stream flow and debris transport, shallow mudflows, rotational slides and rear scarp weathering and mass wastage are the dominant erosion processes. The coastal cliff complex from a point seaward of Gore Cliff (Photo 26) has a long history of instability, with major rockfalls and landslides occurring at regular intervals (Hutchinson, 1987; Bromhead, Chandler and Hutchinson, 1991; Hutchinson, Bromhead and Chandler, 2002; McInnes and Jakeways, 2000; Rust and Gardener, 2002; Rendel Geotechnics, 1994 and 1995a). Nearly 300m of cliff retreat has taken place since approximately 1880 (2.5m per year). At Gore Cliff itself there is strong toe erosion in the Ferruginous Sands and Sandrock, with a recorded mean rate of shoreline recession of 0.6m per year between 1862 and 1980 (Hutchinson, et al. 1981). Preece (1980, 1987) records mollusca and Romano-British artefacts in a chalky hill wash deposit, which he interprets as having been derived by mass movement acting on a slope with a north-easterly aspect. This implies a substantial recession and probable geomorphological re-modelling of the Undercliff in this vicinity over at least two millennia. Alongside the direct contribution of marine erosion to the initiation of slope instability, various other hydrogeological and geotechnical attributes are also of importance (Hutchinson, 1965, 1987; Bromhead, et al., 1991; Rendel Geotechnics, 1994; Barton, 2007). However, the consensus view (Rendel Geotechnics, 1994; McInnes, 2000) is that unloading and over steepening of the cliff base provides the mechanism that prepares for slope failure events that are triggered by periods of intensive or prolonged rainfall adding critically to pore water pressures determined by local groundwater reservoirs (Bromhead, et al., 1991; Hutchinson, et al., 1991; Hutchinson and Bromhead, 2002; Lee, et al., 1998).
The cliffs and platforms of this shoreline are developed in relict landslide debris from successive compound mudslides, rockfalls and multi-rotational slips over several thousand years. The coastal plan is irregular, with a number of confined bays and coves (Posford Duvivier, 1994). During the period for which documentary records are available, parts of the lower Undercliff in this sector appear to have been comparatively stable, with basal recession rates of between zero and 1.0m per year with a space time average of 0.4m per year (Posford Duvivier, 1997). Halcrow (1997) identified relatively slow recession rates of 0.1 to 0.3m per year occurring between 1909 and 1975, but with a significant acceleration thereafter. For the immediate area of St. Catherine’s Point, Hutchinson, et al., (2002) note that recession rates between 1868 and 2001 have been variable, but at a mean of 0.11m per year. Especially rapid recent retreat was noted around the salient of St. Catherine’s Point (Photo 28) and in Reeth (Photo 25) and Binnel Bays where toe erosion has resulted in major upslope reactivations of instability. Periods of relative stability may therefore be interrupted by sharp accelerations in basal cliff recession e.g. 8.75m between 1971 and 1992 at Woody Bay (Rendel, Palmer and Tritton, 1993). Analysis of Coastal Monitoring Programme data showed no discernible net transport between the independent pocket beaches along this section of coast, and the coast protection schemes that have been implemented have reduced potential for toe erosion and slope failure reactivation (McInnes, 2007). For example, the Castle Cove scheme (Photo 29), completed in 1996, has involved slope regrading, provision of a debris accumulation area, groundwater drainage of mudslide sources, a toe loading and wave energy dampening rock armoured revetment and terminal rock breakwaters (HR Wallingford, 1996; Rendel Geotechnics, 1996, 1997; Lee, et al., 1998). The A3055 between St Lawrence and Niton has been realigned over a 300m section following a major landslide re-activation during the winter of 2000/01. The reactivation followed mudslides and failures lower in the undercliffs (Photo 31) and resulted in a period of road closure (Photo 32) following loss of the original road. A stabilisation scheme at Castlehaven, Niton involving application of novel drainage measures to reduce groundwater levels and a rock armour revetment to arrest basal marine erosion was implemented in 2005. The problem involved a sharp acceleration, since January 2004, in the rate of landward reactivation of relic landslides, but remedial work also needed to consider sensitive environmental qualities in developing an appropriate scheme. This particular site, together with many of the others identified above are explained in greater detail by McInnes and Jakeways (2000); Hutchinson and Bromhead (2002); Clark, et al. (2007); McInnes (2007) and Moore, et al. (2007).
The material composing the sea cliffs is made up largely of back tilted landslide debris, dominated by boulder-sized blocks of Upper Greensand, especially the Chert Beds (Photo 14 and Photo 24). There are, however, outcrops of unconcealed Lower Greensand and Gault Clay. The debris apron is at its widest either side of St. Catherine's Point, where borehole data indicates that its seaward portion abuts against a probable buried marine cliff at -7mOD (Hutchinson, 1987; Hutchinson, Bromhead and Chandler, 1991; 2002) which has been tentatively dated at between 7 and 9,000 years BP. After the creation of this cliff line, there was a period of active debris sliding, followed by an interlude of stability represented by tufaceous deposits, and renewed slope wash and intermittent sliding between approximately 5,000-4,500 years BP (Hutchinson, 1987). The series of thick, confluent and possibly superimposed debris aprons provide toe weighting that accounts, in part, for the stability of parts of the Undercliff complex eastwards of Reeth Bay. The seawards extension of these debris aprons is unknown, but there are narrow shore platforms cut across landslide debris at headlands such as Binnel and Woody Points that demonstrate marine abrasion and semi-continuous cliff recession in the period since 4,500 years BP. A recession rate of 0.4m per year has been estimated for the Gault clay cliffs at Castle Cove (Barrett, 1985); this is also quoted for the cliffed coast between St Catherine's Point and Steephill Cove, with a total sediment yield of 115,000m³ˉper year - mostly sand and clay, but with some sandstone boulders, chert, and flints that contribute to the local beaches (Posford Duvivier, 1997). Some of the beach material in the small pocket bays trapped between large blockslides does not appear to be of immediately local origin.
Due to the lack of detailed information in relation to composition of the eroding cliffs it has not been possible to quantify the volume of beach grade material derived from the erosion of the cliffs, although the 2004 proposed rate was estimated at more than 20,000m³ˉper year.
This coastline is developed in a complex of multiple semi-rotational slips, (Geomorphological Services Ltd, 1991; Rendel Geotechnics, 1995; Hutchinson, 1987; 1991; Hutchinson, et al., 1991; Moore, et al., 2007; Moore, et al., 2010; McInnes, 2007) Because of development since the early Victorian period along and immediately inland of much of this frontage, there have been successively more robust and comprehensive schemes to protect the cliff toe (Photo 9 and Photo 10) as documented in the sources given above. Between 1896 and 1974 there was approximately 12.2m of shoreline retreat, releasing somewhat over 500,000 metric tonnes of sediment into the nearshore transport zone (South Wight Borough Council, 1985). Protection measures introduced between 1992 and 2000 have effectively reduced toe recession rates to zero e.g. Photo 6, though cliff-face weathering and shoreface abrasion continue (Posford Duvivier, 1999). Shallow translational slides have been recorded for a number of locations either without protection or where defences have become dilapidated. The role of beach accumulation as a natural defence measure has been debated, and it remains uncertain what proportion derives from local cliff erosion. The renewal of slope instability at the site of Collins Point following beach and debris removal to supply material for the construction of a harbour in 1863 indicates interdependence (Royal Commission on Coast Erosion, 1911). Previously Collins Point had retained a beach of fine shingle supplied by littoral drift, but erosion of the Gault Clay outcrop across the foreshore by wave action resulted in undermining of the newly built seawall. The result of an enquiry was the building of a rock masonry groyne to substitute the former role of Collins Point headland (Posford Duvivier, 1994b). This case example would seem to endorse the view that local beaches, in spite of their relatively small width and volume, are important in regulating rates of coastal slope erosion, as well as adding to the loading of cliff toes.
Another environmental factor that may play a role has been discussed by Hutchinson (1983) and Hutchinson, et al. (1991) with specific reference to the Mirables Undercliff, east of St Catherine's Point (E8). Referring to the zone dominated by mudslides derived from the Gault Clay outcrop, they point to the fact that the critical Gault/Carstone contact is located 1m to 2m above OD, with mean high water spring tides at +1.5mOD and mean low water springs at -1.6mOD. Thus clay outcrops exposed above normal beach levels lie within this tidal range and are subject to the direct mechanical and hydraulic effects of breaking waves. A wave height of 15m once in 50 years has been predicted for this area by numerical modelling (Hydraulics Research, 1990, 1991). Toe erosion may consequently stimulate mudslide activity, the rate of which would be primarily dependent on the efficiency of debris removal. Hutchinson (1983, 1987 and 1991) developed this explanation through comparative analysis of several landslide sites and it may have a relevance to other locations. A factor not taken into account, and which also relates to other sectors of the Undercliff coastline, is the presence of numerous boulders and inter-joint blocks of Upper Greensand scattered across the inter-tidal platform. Derived from previous, more extensive, but now eroded, debris aprons - or possibly from single run-out falls or slides - they constitute a type of nearshore reef that must absorb some incident wave energy. Chandler and Hutchinson (1984) report, using borehole log evidence, that the base of slide-generated aprons extends to approximately -20mOD and rest on an erosional surface cut into the Sandrock. The substantial recession of this cliffline during Holocene times is thus implied, with several alternating cycles of stability and active landsliding. Chandler and Hutchinson (1984) use a radiocarbon date from preserved organic sediments in west Ventnor to indicate general stability between 4,000 and 4,500 BP, with renewed activity between 4,000 and 3,500 BP. A similar chronology is proposed by Preece, et al., (1995) with reference to colluvial infill in a dry valley near Ventnor.
The rates of erosion reported for this coastline all refer to the situation predating the installation of localised seawall, groyne and breakwater protection systems. They vary from 0.15m per year for the Bonchurch frontage, 0.3m per year between Steephill and St. Catherine's (McInnes, 1994) to less than 0.05m per year in the Upper Greensand boulder-dominated debris aprons immediately west of the Western Esplanade, Ventnor (Posford Duvivier, 1991, 1993c; McInnes, 1994). Rendel Geotechnics (1993) suggest 0.1m per year for Woody Bay (E8) but with considerable temporal variation affecting both the cliff base and cliff top. A cyclical behaviour of alternating periods of rapid and comparatively slow erosion is tentatively proposed. The western cliffs at Ventnor have been protected by a rock revetment constructed at the cliff toe in 1992 (Photo 30). In the early 1990s a massive, stepped sea wall was built along the length of Wheeler's Bay to replace pre-existing sloping timber defences (Photo 9 and Photo 10). The latter, completed in 1973, proved ineffective in trapping sufficient beach shingle and was breached on several occasions by wave action. An 80m section of the seawall at Wheelers Bay was strengthened in 2000 by the addition of a seaward buttress, soil nailing, drainage and grass matting of the rear slopes (Posford Duvivier and Malcolm Woodruff, 1998; McInnes and Jakeways, 2000; McInnes, 2007). Sediment by erosion is now restricted to wave scour of the foreshore platform (Posford Duvivier, 1999). The toe loading effect of these structures should also enhance the stability of the former sea cliffs. Monk's Bay at Bonchurch was formerly protected by a partial sea wall and concrete groynes; Posford Duvivier (1990b) calculated an erosion rate here of 0.2 to 0.3m per year, with up to 15m of cliff recession since 1970 where the sea wall had virtually collapsed. During the winter of 1989/90, a semi-rotational landslip released 100,000m³ˉof material, but slope stability has been improved since the completion in 1992 of a comprehensive protection scheme incorporating offshore rock breakwaters, 30,000 tonnes of shingle beach replenishment, slope reprofiling and drainage (Posford Duvivier, 1992). Consequently, sediment input from this source has been substantially reduced (Photo 6).
Due to the lack of detailed information in relation to composition of the eroding cliffs it has not been possible to quantify the volume of beach grade material derived from the erosion of the cliffs, although the 2004 proposed rate was estimated at more than 20,000m³ˉper year.
At Dunnose, there is a sharp change in coastal orientation, and the cliffs in this area are undefended and eroding. Cliff height and form is initially similar to the immediate west of Ventnor. The 30 to 40m high sea cliffs in the south part of this sector occupy a cross-sectional area of the eastern sector of the Undercliff cut into landslide debris, gently dipping Gault Clay and Upper Greensand and Sandrock (Lower Greensand) further north. On the eastern margin of Bonchurch major landslides occurred in 1810, 1818, 1904 and 1910 with significant reactivated rotational failure events in 1961/2, 1988, 1995 and 2001 (Palmer, et al., 2007). South of Luccombe Chine there is evidence for marine erosion of the cliff base, although translational slides and mudflows are frequent and often temporarily conceal bedrock before their removal by marine erosion. Terrace recession has exposed mudslide and landslide material within the Gault Clay, resulting in some reactivation of cliff top retreat. Mudslides move across successively lower Greensand benches, where they are contained as temporary stores. Cliff profiles assume a more degraded form where there are substantial accumulations of boulders across the foreshore. These derive from fresh falls, slides and toppling failures and the removal of less resistant clays and sands within landslip debris aprons created by previous major landslips. Coastal Monitoring Programme data indicates that locations of potential cliff input are restricted to episodic landslides or short lengths of cliff, with significant erosion of the cliff or slope. The apparent limited accretion at the toe indicates that fine grained sediments are not retained on the foreshore. The cliffs north of Luccombe Chine assume a more complex composite form with distinct benches. The area directly inland of Luccombe Chine has a well-documented history of part translational and part rotational slope failure (Geomorphological Services, Ltd., 1989). The re-exposure of failure planes in both the Gault Clay and consolidated landslip material by toe erosion may have been a trigger to earlier major landslide events, although groundwater conditions - notably critical pore water pressures induced by exceptional rainfall - are also important (Palmer, et al., 2007). The 1995 event, involving failure of the rear scarp and retreat of the cliff top, displaced approximately 800,000m²ˉof rock debris and its deposition on the adjacent beach and foreshore (Rendel Geotechnics, 1995b; Palmer, et al., 2007). The 2001 event affected much of the same area but caused failure of the Upper Greensand bench and upslope regression of the head scarp. It occurred within a short period of intense winter rainfall, when 530mm of effective precipitation was received in three months. Despite a wide inter-tidal sandy beach (over 100m at maximum spring tides) at the mouth of Luccombe Chine, basal cliff trimming and notching by waves is an active process. Coastal Monitoring Programme data indicates that Luccombe beach shows an erosive trend, with material transported northwards on the sub-tidal foreshore, with Horse Ledges acting as a partial littoral boundary. Debris loading of the benches by landslip debris from above is associated with groundwater seepage at the junctions between interbedded sandstones and clays. Water supply and waste water leakage from Luccombe village following the provision of these services in 1927 was considered to be an additional contributory factor (Geomorphological Services, 1989; Watson and Bromhead, 2007). Breaking waves of a probable maximum height of 3m could be propagated across an easterly fetch of up to 170km, and might generate rip currents of sufficient velocity to rapidly remove a proportion of the fine-grained fraction of cliff fall debris and beach sediments (Geomorphological Services, 1989).
Below the Luccombe road, for a short distance just south of Shanklin Chine, some timber structures are in place reducing the rate of cliff erosion. There are no estimates of rates of erosion in this coastal sector based on experimental data; the few figures quoted in the literature rely upon analysis of the position of the cliff top on successive editions of Ordnance Survey maps, the accuracy of some being in doubt. Barrett (1985) and Halcrow (1997) quote a rate of recession of 0.2-0.3m per year at Dunnose, Posford Duvivier (1981) calculated a retreat rate of 0.3m per year and Posford Duvivier (1987) suggest up to 0.5m per year for this length of shoreline. Based on more detailed investigation, Palmer, et al., (2007) suggest a retreat rate of 1 to 2mm per year for the past two centuries. This may not take full account of the loss of temporary stores of landslip material, and it undoubtedly generalises significant spatial variation, e.g. a rate of 1.0 to 2.5m per year in the vicinity of Borderwood Lodge (Halcrow, 1997). A mean recession rate of 0.4m per year gives a total potential sediment yield of 75,000m³ˉper year (Posford Duvivier, 1999), however, analysis of Coastal Monitoring Programme data suggests that this volume is an overestimate, and a large proportion of available material would be fine-grained and not readily retained on the foreshore. At Luccombe, this latter figure may be a close approximation to the recession rate maintained over the past 140-150 years (Posford Duvivier, 1990b), but is probably exceeded along the shoreline between Yellow and Horse Ledges. McInnes (1994) proposes cliff retreat at a rate of 0.2 to 0.3m per year between Appley Steps and Steed Bay. North of Horse Ledge, cliff foot erosion is inhibited to some extent by partly redundant groynes, and the scree that has accumulated in front of Knock Cliff and Apley Steps indicates that small scale but frequent rockfalls and toppling failures, due to weathering and stress relief, are significant. The general morphology of the near-vertical cliff line between Yellow Ledge and Shanklin Chine indicates the longer term dominance of block failure and bench formation associated with aquicludes. Here, and also at Luccombe (Geomorphological Services, Ltd., 1989; Moore, et al., 1991), the rate of cliff top recession, at 0.3m per year, appears to have accelerated over the past century, inducing failure reactivation on several occasions (Posford Duvivier, 1990).
Almost the entire frontage is made up of substantial sea walls and groynes, so that the former cliffs are no longer subject to marine erosion and cannot make a contribution to sediment supply. The history of protection can be traced back to the 1830s and is fully documented (Lewis and Duvivier, 1974; Posford Duvivier, 1981, 1989a, 1993a). All but one of the sea wall sections were finally joined together in 1934. Barrett (1985) quotes a figure of 0.2 to 0.4m per year retreat for the Littlestairs cliffs prior to their protection in the mid-1970s. Assuming that there has been little subsequent change in cliff height, this rate would yield a former supply of about 3,000mm³ of sand per annum. The effect of this loss of sediment feed could be reflected in the estimate of foreshore lowering of 0.3m per year between 1960 and 1970 (Barrett, 1985). This figure appears to have been derived from surveys of the Lower Greensand shore platform exposed seawards of Lake at low spring tides. It is part of a feature that is now fully exposed in front of the sandy beach around much of the perimeter of Sandown Bay, possibly pointing to long term reduction of beach sediment storage. Although isolated from wave activity, the former 40m high sea cliffs remain geomorphologically active, due to sub-aerial weathering and mass movement (Rendel, Palmer and Tritton, 1988). Barton (1985, 1991) estimates 5m of slope crest recession between 1907 and 1980 (0.06m per year), for part of the cliffline between Shanklin Chine and Hope Road, a figure derived from measurement of the dimensions of basal scree deposits. Various protection techniques including cliff-top regrading, drainage, timber shuttering, geofabric/grass matting, netting, rock bolting and talus reprofiling and removal have been implemented to manage this problem (Clark, et al., 1993; Rendel Geotechnics, 1991b, 1992; McInnes, 2000; Royal Haskoning, 2010) over a 3.5km frontage at Shanklin. This has not been entirely successful with several small free face detachments and a major talus slope failure in March 2001 that released 5,000m³ˉof material following intense rainfall and a sustained period of easterly winds; there was further failure in early 2014 in response to exceptional precipitation.
Immediately north-east of Yaverland the seawall terminates and there is no northwards protection against marine erosion. The outcropping strata are lithologically varied but collectively unresistant, excepting the Chalk. Shallow translational slides and mudflows are characteristic of the Wealden shales and clays, and give an irregular low, cliff profile that frequently exhibits basal notching. The Ferruginous Sandstone of Red Cliff is comparatively more coherent and supports a near vertical lower cliff face. Multiple translational sliding and mudslide surging in the Atherfield Clay has created a 130m wide degradation zone that has been semi-continuously active since at least 1910 (Hutchinson, 1965) whilst both deeper seated and superficial mass movement affect the Gault Clay outcrop, giving a retreat rate of 0.5m per year (Halcrow, 1997). The rate of erosion in the Wealden Marls at Yaverland was calculated to be 60m between 1910 and 1945 (Lewis and Duvivier, 1973), with annual rates varying between 0.5 and 2m per year, depending on beach width. Foreshore recession of 0.3m per year, 1896-1969 is indicated from Ordnance Survey maps (Posford Duvivier, 1981). A figure of 0.35m per year, for the coastline south of the Chalk outcrop, is contained in an analysis of all historical sources (Posford Duvivier, 1997; Posford Duvivier and British Geological Survey, 1999). Lithological changes impose spatial variation of recession rates of change between 0.2 and 0.8 m per year. The foundations of early nineteenth century buildings at Yaverland Fort, now exposed on the foreshore, indicate 0.5km of cliffline retreat, and a similar rate on the Gault Clay outcrop between 1870 and 1980 (Barrett, 1985). Repeated semi-rotational slides, and their rapid removal by wave action, have resulted in as much as 20m of cliff top retreat in less than one year at specific sites (Barrett, 1985) with instability evident up to 70m inland. Hutchinson (1965) reported that the remains of a seawall built in 1924 to protect Yaverland Castle was 80m seawards of the cliffline in 1964. There are no reliable records of shoreline change along the south facing Chalk cliffline seawards to Culver Cliff, but basal accumulation of boulders, shallow slides, rockfalls and talus cones, indicate the contemporary, as well as long-term, effectiveness of both marine and sub-aerial denudation. Posford Duvivier (1999) propose an overall recession rate of 0.23m per year, but it is not clear how this figure was calculated. McInnes (1994) calculates retreat at about 0.1m per year. Active erosion is evidenced by caves, incipient stack formation and a well-defined shore platform at Whitecliff Ledges northwards to the blunt headland defining Whitecliff Bay. Taking average cliff height and the above quoted erosion rate, Posford Duvivier (1997) suggest that the Chalk cliffs yield some 30,000m³ per year of material, of which only about 2% (<500m³) consists of flint. Analysis of the Coastal Monitoring Programme data supports less than 1,000m³ per year of beach grade material is derived from the erosion of the cliffs which is then available for northeastward transport, with clays and fine-grained sediment not retained on the foreshore, being transported into the nearshore; this volume is a reduction from the 2004 proposed rate of 3-20,000m³ per year.
In addition to sediment losses from cliff erosion, abrasion and scour of the intertidal shoreface also contributes input to the littoral transport system. A range of figures are presented in Posford Duvivier (1999), all of them derived from the application of a basic methodology to local conditions of shoreface width, water depth and rock erodibility. Rates vary from 2 to 35mm per year of vertical corrasion, yielding from less than 500 to over 24,000m³ˉper year of sediment. Quantities are largest where sandy or clayey rocks form the shoreface substrate on the West and South coasts. It is presumed that almost all of this sediment is fine grained and this therefore not retained local beaches.
3. Littoral Transport
Introduction
It is assumed that sediment transport is principally wave-driven, as tidal current velocities are mostly low (0.3-0.5ms-¹). A large proportion of the fine sediment released by erosion of sandstone, clay and marl rock units is not retained by local beaches, and ultimately moves offshore, in suspension (Posford Duvivier, 1999; Brampton, et al., 1998). As a general trend, beaches consist of a relatively narrow and in places impersistent gravel backshore and a wider sandy foreshore, which extends below MLWS; they progressively steepen between Freshwater Bay and Rocken End. South to Atherfield Point beaches are composed of material derived from local cliff erosion, but south of this location an offshore supply source makes an additional contribution. The gravel component becomes more dominant, although the median grain size of coarse clastic material gets smaller, in a south-eastwards direction.
The presence of a fringe of coarse gravel and boulders accumulations derived from cliff block falls has been discussed previously. The 2011 Coastal Monitoring Programme swath bathymetry data, between the Needles and Afton Down, shows the southerly extension of the cliff toe and sediment free rock platform outcrops offshore approximately 500m, maintaining the west-east sinuous alignment of the cliffs. There is no evidence of along-shore sediment transport between Scratchall’s and Freshwater or Compton Bays. Further offshore and extending east to Compton Bay the marine sediments are sufficiently thick to cover the underlying bedrock.
The largest beach accumulation is at Freshwater Bay, where it is characterised by multiple berms and a steep backshore storm ridge, banked up against the cliffs throughout most of the year, particularly in the eastern sector. The height and continuity of this feature led Barrett (1985) to suggest that this beach is strongly reflective. The material is composed predominantly of large, sub-rounded to sub-angular flint clasts, suggesting that abrasion by swash and backwash reworks sediment trapped in this well-defined coastal inset. The lack of in situ flints in the Chalk cliffs in the eastern part of the Bay suggests their movement by littoral transport from the west, but there may be an input from the mass wasting and marine erosion of the soliflucted chalky-flint deposits infilling the truncated valley profile of the Yar. This would have been more significant before the completion of the first generation of sea defences in the late nineteenth century. Severe damage sustained by the sea wall esplanade and groynes, necessitating extensive repairs and reconstruction in the 1900s, 1953 and 1966, indicate the effectiveness of both abrasion and scour (Posford Duvivier, 1989a). The consensus view is that the small tidal range helps to construct and conserve an essentially stable beach, which now offers adequate protection to the foundations of the seawall, except perhaps for the salient around the Albion Hotel (Posford Duvivier, 1989b). The esplanade sea wall built in the late 1890s (extended and rebuilt in 1951), was constructed on the foundations of a backshore berm, thus immobilising a previously dynamic sediment store. Coastal Monitoring Programme data supports no net longshore transport within Freshwater Bay, as the spatial and temporal variability is driven by and responds to wave direction and conditions. Colenutt (1904) reported that short-term and short distance east to west movement had been observed following south-easterly winds. The beach to the immediate west of the Albion Hotel was depleted in 1902-4, possibly due to gravel trapping by groynes or deliberate shingle removal for construction materials. All subsequent descriptions of this beach fail to mention any patterns of sediment transport reversal or net accumulation/depletion. The groynes, present since the 1930s, have tended to minimise changes in beach form and orientation. The cessation of beach quarrying circa. 1910 may have been significant in restoring dynamic equilibrium after an unknown period of selective removal of clasts.
This long sector of coastline has a north-west to south-east trend, except for the stretch between Freshwater Bay and Compton Down where the Chalk cliffs and boulder beach have a similar character as the coastline to the west, although the beach is marginally wider. Chalk from the eroding cliffs is transported southwards towards Hanover Point, but it is virtually absent beyond 1km off Freshwater Cliff (May, 2003). The change in orientation at Compton Chine reflects the influence of rock type and geological structure and coincides with a marked change in beach character. The 2011 Coastal Monitoring Programme swath bathymetry data, between Freshwater Bay and Rocken End is characterised by extensive areas of sediment-free, exposed bedrock, extending from the toe of the cliffs and inter-tidal rock platforms. As a generalisation, shingle forms the upper backshore, with fine to medium sand constituting the mid and foreshore areas. Under high energy wave conditions, this sorting pattern is often lost, and beach composition is mixed. The sand derives mostly from the long term erosion of the cliffline and of substrate silts, clays and shales exposed across the foreshore. Fine-grained sediments are transported offshore by suspension (Posford Duvivier, 1999a; Brampton, et al., 1998). The beaches between Compton Bay and Atherfield Point are of moderate gradient, and the inter-tidal zone is wide. Nearshore significant wave heights vary from 1.6 to 7.5m, the latter with an approximate 1 in 100 year return frequency (HR Wallingford, 1999). In Chale Bay, the dominantly flint gravel beach is substantially higher, steeper and comparatively narrow, about 30m at mid-tide in the vicinity of Whale Chine compared to 70m at Compton Bay (Barrett, 1985). Rock ledges and platforms are a further feature of the inter-tidal zone between Brookgreen, Hanover Point and Atherfield Point which help to dissipate wave energy and conserve beach form. Beach profiles and volumes are subject to substantial fluctuation, given the potential range of wave heights and periods (HR Wallingford, 1992). Large rafts of seaweed are frequently superimposed on the mid- to back-shore beach, which act as a buffer against wave scour, particularly during late spring and early summer. Norman (1887) described the southerly beaches attaining "great depth", and composed of medium to fine gravel. It was his unsubstantiated view that the material derived from offshore deposits, reworked by wave action.
Colenutt (1938) reports that the normally coarse sandy beach at Brookgreen had been "occasionally" replaced by well graded, polished (abraided?) gravel, but there are no other accounts of short term changes in beach sedimentology except for a tentative inference by Kay (1969) that the pebbles composing Chale beach had recently changed from large to smaller size. The same author notes that this beach may have been steeper in the late nineteenth century, or that a permanent storm ridge might have existed at this time. It would appear to be rare for a true storm berm to form at most points along this coastline. If it does, it only persists for very short periods.
Coastal Monitoring Programme data supports the dominant littoral transport direction is from north-west to south-east (Colenutt, 1938; Kay, 1969; Posford Duvivier, 1981, 1989a, 1999; Barrett, 1985; Brampton, et al., 1998), with updrift accretion occurring at salient features, such as Hanover Point. The reduction in the mean size of gravel from north-west to south-east is also consistent with progressive abrasion along this pathway (Barrett, 1985), although in itself it is not a reliable guide to dominant drift direction. Halcrow (1997) suggest that the coarse fraction, whilst retained on the inter-tidal slope, moves downdrift rapidly. Beach material therefore has little long-term stability, and may therefore fail to provide effective cliff toe protection. The relatively limited development of beach width in Chale Bay has not been satisfactorily explained, (Kay, 1969). Barrett (1985) has proposed that the headlands confine the littoral transport of gravel, but as they are relatively minor salients this is a debateable proposition. Coastal Monitoring Programme data also indicates no direct evidence of offshore losses, nor of net offshore to onshore feed of gravel (Posford Duvivier and British Geological Survey, 1999). It is unlikely that more than a small fraction of the gravel constituting the beach in Chale Bay is derived from the erosion of the Chalk cliffs of Freshwater Bay, via longshore drift. Although there are cappings of valley gravels at the cliff top between Shippard's Chine and south of Brookgreen, they are thin and patchy and even long-term cliff erosion would not generate a sufficient quantity of beach sediment from this source. An off to onshore wave-transported input of gravel is therefore a reasonable inference (Brampton, et al., 1998). The dynamics of longshore and nearshore sediment transport for much of the western Wight coastline are therefore not understood. What little evidence that is available is based on chance observations. The important interacting variables controlling transport are offshore bathymetry, wave climate and the nature of the shallow seabed sediments potentially available for onshore transport (Brampton, et al., 1998). Davies (1997) has suggested the possibility that there are discrete sediment transport sub-cells along this length of coastline. When waves propagating from the South-West are operative, there is only minimal exchange between them; however with incident refracted waves approaching from the East or South-East, net alongshore westwards transport takes place. This could introduce a source of gravel from a nearshore or offshore source, with finer grained sediment removed by suspension transport. The only characteristic that is not in doubt is the change from dissipative (sand) to reflective (gravel) beaches from north-west to south-eastward and the progressive reduction in the mean size of the gravel fraction. Roundness may also increase in this direction.
The 2011 Coastal Monitoring Programme swath bathymetry data, between Rocken Point southwards and around St Catherine’s Point, is characterised by relatively minimal sediment, rock platforms and exposed outcrops. The nearshore seabed between St Catherine’s Point to Monks Bay is a complex mix of rock platform and outcrops close to shore and rock covered in varying thicknesses of sediment further offshore. Small constrained patches of sand and coarse sediment are evident in numerous pocket beaches between St Catherine’s Point and Luccombe Bay including Reeth Bay, Ventnor Bay and Monks Bay. Offshore of Ventnor Bay there is an area of large sedimentary bedforms, the majority of which are slightly asymmetrical indicating west to east sediment transport flow, although further offshore are smaller-scale slightly asymmetrical bedforms that indicate east to west transport.
Rocken End, made up of landslip debris, retains a beach composed of fine gravel described as "pea" gravel (Barrett, 1985; Posford Duvivier, 1981; Rendel, Palmer and Tritton, 1993; Halcrow, 1997). The latter is characterised by well-sorted, sub-angular to sub-rounded flint clasts of a mean diameter of 10mm (range of 6 to 15mm). It is usually highly polished, indicating abrasion. The transition from coarse to fine gravel between Walpen Chine and Rocken End is striking, but defies ready explanation- but see Davies (1997), discussed in the previous section, for a possible insight. Narrow beach widths may be due to the updrift interception of littoral drift. Between St. Catherine's Point and Ventnor there are several well-defined pocket beaches of similar "pea" gravel (Photo 5, Photo 14 and Photo 24); however, Coastal Monitoring Programme data indicates no direct evidence of material by-passing St. Catherine's Point. Barrett (1985) has observed that these pocket beaches are adjusted to incident wave approach and exhibit weak west to east littoral drift, as supported from Coastal Monitoring Programme data. Brampton, et al., (1998) suggest that some 45% of nearshore waves approach from the west or south-west, and 28% from east or south-east. There appears to be little exchange between adjacent bays, but by-passing may take place when there are oblique south-westerly long period waves. Some beaches, particularly at the eastern end of this coastline, have been subject to draw down, indicating that potential rates of transport exceed available supply. Tidal currents may play a minor role in moving finer grained material, particularly in the vicinity of St. Catherine's Point where velocities are at a maximum. Posford Duvivier (1989a) record that attempts in the 1970s to accrete beach material along Ventnor's Eastern Esplanade (Photo 5), using groynes, were unsuccessful. In this case, however, wave abrasion of the concrete seawall was as important a cause of this failure as the difficulty of retaining a permanent beach of sufficient height and volume to absorb wave energy. Groynes and breakwaters elsewhere, e.g. Swale and Linnington groynes at the Western Esplanade, indicate net west-to-east littoral transport, but the ultimate source of the "pea" gravel that they retain is uncertain. Posford Duvivier (1981, 1989a) infer that it is derived from Chale Bay, but in view of the evidence of negligible lateral transfer from one bay, or cove, to the next (i.e. a sequence of small independent sub-cells) a pathway of shore parallel movement that bypasses Rocken End and then moves onshore further east is conceivable (Rendel, Tritton and Palmer, 1993; Brampton, et al., 1998). Nearshore sediment sampling has revealed the presence of sandwaves, between 400 and 900m offshore southwest of Ventnor; sand patches and ribbons are also present, suggesting inshore movement of sand between more stable gravel deposits. Net transport of sand offshore appears to be west to east (Rendel, Palmer and Tritton, 1993).
In some of the bays, e.g. Reeth Bay (Photo 25), sand is co-dominant with gravel, and elsewhere coarse sand patches alternate with fine gravel under varying wave conditions. It is therefore probable that sand provides a foundation for a comparatively thin veneer of gravel, which is apparently quite mobile (Barrett, 1985). Norman (1887) described a beach "near Ventnor" as "covered with ... sand ... twenty years ago", and implied that there was no gravel along the entire southern coastline before the 1840s. Boulder-sized Upper Greensand blocks constitute the beach frontage wherever sand or gravel is absent, having been released by the weathering and erosion of landslide debris. The sand fraction probably derives from the same source. Littoral sediment volumes are temporarily augmented by sudden inputs from slides or slumps, although the only documented example is the rapid break-up and disruption of a rockfall or slide cone at Rocken End in 1825 (Barrett, 1985). Immediately east of Ventnor, particularly between Bonchurch and Dunnose, beach sediments wholly derive from landslip debris and include coarse gravel from the flint and chert content of fallen blocks of Chalk and Upper Greensand respectively. Coastal protection schemes throughout this century have progressively reduced inputs from cliff erosion, so that offshore to onshore transport may be an intermittent source of supply to beaches. Nourishment has also been undertaken since the mid-1970's e.g. Monk's Bay (Photo 6). Between Western Esplanade and Wheelers Bay the completion of massive seawall protection in 1988 (Photo 9 and Photo 10) now ensures no contemporary contribution from basal cliff erosion. Severe downdrift reduction in sediment supply (as evidenced by the undermining of groynes) may have been a cause of accelerated erosion in Monk's Bay between 1950 and 1990 (Posford Duvivier, 1981). The lack of any eastwards transport of "pea" gravel beyond Ventnor Bay has attracted some comment (e.g. Posford Duvivier, 1981), but the reasons for this are obscure. It does, however, suggest that the littoral transport system is compartmentalised, perhaps due to complex patterns of near and offshore submarine relief related to submerged landslide debris.
The 2011 Coastal Monitoring Programme swath bathymetry data indicates that the seabed between Luccombe Bay and Culver Cliff is predominately covered with sufficient thickness of sediment to mask the underlying bedrock. Rock ledges and isolated outcrops occur offshore of the headlands at Dunnose, Horse Ledge, Shanklin Groynes and Chine and Culver Cliff.
Coastal Monitoring Programme data indicates a net northward littoral drift from Luccombe Bay towards Shanklin, with accretion against Horse Ledges, and an erosive trend further north towards Shanklin Groynes. Much of the backshore and inter-tidal zone, which includes rock-cut platforms, is littered with large Upper Greensand and Lower Greensand sandstone boulders. The main exception is Luccombe Bay, where a wide, fine-grained sandy foreshore has accumulated. There are, however, distinct pockets of coarse, angular chert and flint particles at the cliff base between Dunnose and Luccombe confined between salients of both bedrock and residual landslide debris lobes (Photo 3 and Photo 16). Between Yellow Ledge and Shanklin Chine this coarse material becomes a more defined backshore berm. Both Yellow and Horse Ledges intercept the littoral transport of sand, but are comparatively easily by-passed if the rapid extension in the width of sandy foreshores immediately north of both of these features are reliable indicators. The principal features of the beach and foreshore of this sector have not changed appreciably in recent decades if the description by Colenutt (1938) is to be relied upon.
The beach along this frontage comprises homogeneous sand although the zone immediately east (downdrift) of Yaverland changes to sand with patchy gravel. There are clearly defined offsets in beach width associated with the numerous groynes, which indicate that the dominant longshore transport is from south to north (Photo 7, Photo 12 and Photo 13). With the change in coastline orientation east of Yaverland, and the effect of Culver Cliff on wave refraction, an east to west counter-drift may operate at times when waves from an easterly or south-easterly fetch prevail. The gravel beach between Red Cliff and Culver (Photo 37 and Photo 39) is field evidence in support of this, although an offshore supply source, or a possible barrier origin, cannot be discounted. Occasional severe draw down of the sandy beaches at Shanklin and Sandown has revealed a shingle basement (Lewis and Duvivier, 1973), most probably derived from transport from the south.
The 2011 Coastal Monitoring Programme swath bathymetry data indicates that the seabed between Luccombe Bay and Culver Cliff is predominately covered with sufficient thickness of sediment to mask the underlying bedrock. Rock outcrops can be seen below a thinner veneer of sediment in various areas throughout Sandown Bay. A few localised areas of bedforms are discernible in the east of Sandown Bay, potentially formed in response to tidal current velocities and directions associated with rock ridge outcrops south of Culver Cliff.
The whole of this frontage has been subject to intensive protection, with successively more comprehensive measures incorporating seawalls and groynes installed since the 1860s. Unusually complete and detailed records of shoreline management have survived, and are summarised in several consultants' reports (Lewis and Duvivier, 1973, 1981; Posford Duvivier, 1989a, 1990b, 1993b, 1995; Halcrow, 1997). These documents record the length, height and date of construction of all the major groynes and provide some measurements of volumes of sand retained updrift. As an example, Herne Hill groyne at Sandown, established in 1860 and rebuilt in 1893, accumulated sufficient sand to promote a 12m seaward advance of the mean High Water Mark and provide the foundation for the forward building of the Esplanade between the 1890s and 1930s (Photo 40). Small Hope groyne, at Shanklin, built in 1860 and rebuilt in 1901, created a downdrift offset of 60m and a "step" of 15m (Photo 7). South of Shanklin Chine there has been some 12m of seaward migration of Mean High Water since 1896. There have been many instances where inter-groyne sectors of the beach have suffered depletion (e.g. Photo 34) and where Mean High Water has advanced landwards and the beach has steepened in response (Photo 8, Photo 41 and Photo 42); in one case, in northern Sandown, the rate of movement was approximately 30m in as many years in the period 1920-1960 (Posford Duvivier, 1989a). The negative downdrift effects of groyne construction between 1890 and 1950 is a classic example of an effect experienced in numerous other locations on the south coast of England (Barrett, 1985; Halcrow, 1997; McInnes, 1994). Because of the arcuate shape of Sandown Bay, the rate of littoral transport diminishes northwards in response to a reduction in the obliquity of angle of wave front approach. Volumes of littoral drift also diminish along this unit, partly or substantially because of increasing distance from supply sources. The long-term problems of retaining a wide and stable beach have therefore been greater in the northern part of this sector. It is probable that volumes of sediment moving downdrift also decrease in the same direction as a consequence of storage in groyne bays. Supply deficit is also a consequence of the removal of sediment supply from cliff erosion as a direct result of seawall/esplanade construction. With the exception of the Littlestairs section (up to 1974), all of this coastline has been "walled up" since the late 1940s. The only sources of natural replenishment now available are from the erosion of the sandstone cliffs south of Appley Steps, and - possibly - from offshore. It is arguable whether refracted swell waves would disturb this material, and tidal velocities are inadequate to account for significant offshore to onshore transport. If there is a tidal component in the sediment budget of Sandown Bay, it is probably represented by net offshore movement of suspended sediment (Dyer, 1972, 1980, 1985; Halcrow, 1997; Posford Duvivier, 1999). Analysis of Coastal Monitoring Programme baseline topographic (2004-12), lidar (2006-12) and aerial photography (2001/08/13) data confirms a net northeastward littoral along-shore drift with rates of movement 1-3,000m³ˉper year for this frontage, with minimal cliff inputs, although volumes decrease eastwards at Culver cliff. The wide sub-tidal extent of the beach suggests cross-shore transfers may be more dominant, and probable linkages with nearshore and offshore sources of sediment input. The bi-annual frequency of the topographic and lidar data is insufficient to be able to reliably infer seasonal beach changes. From the existing survey data, there does not appear to be a systematic variation in beach levels, width or extent; patchy erosion and/or accretion is evident along the whole Sandown frontage. Although still negligible, there are slightly higher levels of erosion between Sandown Pier and Shanklin than further north along the undefended coast at Yaverland.
Surveys and observations carried out in the 1980s (Posford Duvivier, 1989a) appear to indicate that some of the inter-groyne beaches have stabilised and achieved an equilibrium condition adjusted to the current volumes of input and output. In others, the absence of a permanent backshore shingle berm has promoted wave reflection from seawalls, and therefore beach drawdown (Photo 11, Photo 41 and Photo 42). Some complications to sediment grading resulted from beach cleansing operations during the 1980s and 1990s, but the practice of removing coarse material has now been discontinued.
Littoral drift at the northern end of this sector may have contributed to the transport pathway north of Culver Cliff, before its emergence during the Holocene transgression as a barrier to movement. Dyer (1985) notes that the mineralogy of Ryde Sands bears some comparison to that of the Lower Greensand in Sandown Bay. However, erosional scour of rock outcrops over outer areas of the bay may be a supply source, and this may continue to operate.
4. Sediment Outputs
4.1 Offshore Transport
O1
A map of postulated sand transport pathways around the Isle of Wight (Dyer, 1985), indicates a net onshore to offshore movement away from the approximate location of Hanover Point and possibly seawards of Freshwater Bay. Interpretation and analysis of the 2011 Coastal Monitoring Programme swath bathymetry data, clearly shows the high wave energy nearshore zone of the southwest Isle of Wight is characterised by extensive areas of exposed bedrock, extending from the toe of the cliffs and inter-tidal rock platforms. These rock outcrops in the nearshore and sub-tidal zone are interspersed with constrained areas of sand and sandy gravels, which are largely of insufficient thickness to mask the underlying bedrock. The 2004 arrows have been removed as there is no conclusive evidence to support offshore transport in the southwest or southeast nearshore areas. Where accumulations of sand and mixed sediment occur these are largely constrained within the sub-tidal rock platform. Between Atherfield Point and Rocken End the extent and thickness of sediment, which extends perpendicular offshore from the MLW, may indicate an offshore pathway for sand and mixed sediment, although further monitoring is required to verify. From Rocken Point southwards and around St Catherine’s Point the rock platforms and exposed outcrops are more evident.
4.2 Beach Mining
There are three sites for which there are historical records of the removal of beach sediments for construction materials, as follows:
Freshwater Bay
Colenutt (1904) provides a brief account of an apparently long history of unregulated removal of gravel and states that: "many hundreds of tons of shingle have from time to time been carried away." It was used for concrete in the building of local forts and batteries and the seawall/esplanade in the late nineteenth century. It was the view of Colenutt (and others) that this activity was the principal cause of accelerated beach and cliff erosion in the first decade of the twentieth century; he records 0.8m of cliff recession in the western part of the bay during the winter of 1903/4 and the destruction of the esplanade by "damaging gales" over several successive winters. It is, of course, very possible if not probable that the construction of the esplanade promoted onshore to offshore drawdown of the beach due to reflective waves. Colenutt also reports an absence of sediment accumulation "below high water" in 1902. There is no record of when this activity ceased, but Colenutt's paper proved to be an effective articulation of local concern and it is probable that gravel removal had ceased by 1910 or thereabouts.
Collins Point
The removal of all beach material for concrete aggregate used to construct harbour breakwaters occurred in the mid-1860s. The acceleration of cliff instability was an almost immediate consequence and caused considerable public concern and resulting official remedial action in the mid-1870s.
Sandown and Shanklin
Lewis and Duvivier (1974) refer to selective "pebble" removal from Sandown and Shanklin beaches for amenity purposes. It is uncertain when this commenced but it presumably refers to the removal of gravel patches overlying sand. It is, however, known to have been discontinued in the mid-1990s. Patches of gravels, spread as a thin veneer, do occasionally accumulate on the mid-beach area of Sandown Bay after winter storms but appear to be fairly rapidly removed either offshore or alongshore.
4.3 Dredging
As elsewhere along the south coast of England, there has been speculation that offshore aggregate dredging might be an independent cause of beach depletion. Reliable data on offshore to onshore transport is too limited to either confirm or deny this possibility, despite some strong assertions. Marine aggregate removal has taken place for more than 30 years in a number of licensed blocks off the western, south-western, south-eastern and eastern coastlines. Combined extraction rates increased from 1,143,750m³ˉin 1982 to 1,500,000m³ˉin 1987. An unusual feature is that the Isle of Wight Council has licensing rights purchased from the Crown Estate Commissioners in 1949 by a predecessor authority. The consensus view (Hydraulics Research, 1984; HR Wallingford, 1992) is that coarse sediment in water depths exceeding 15m is effectively immobile, and thus does not contribute to the littoral sediment budget.
Pot Bank
The permitted dredged area is slightly more than 754,000m²ˉand is centred about 1100m due south of the western extremity of the Bridge, a submarine ledge extending westwards of the Needles. Dredging commenced in the late 1940s and between 1950 and 1980 a total of approximately 5,062,500m³ˉof gravel was removed. During the 1980s and early 1990s quantities have been much smaller, amounting to less than 315,000m³ˉbetween 1982 and 1987. Licenced extraction was suspended in 1994. A detailed analysis of seabed configuration between 1950 and 1970 using hydrographic chart data (Geodata Unit, 1989) revealed fairly modest seabed changes and concluded that some gravel may be moving into the area to partially replenish that lost to dredging. As Pot Bank is 7km from Freshwater Bay, this is an unlikely supply pathway and its small size and limited height even before dredging would have a limited impact on wave refraction patterns. Moreover seabed changes during the period of most active dredging were no greater than those recorded on pre-dredging hydrographic charts between 1937 and 1951. Dyer (1985) indicates, without direct reference to the supporting evidence, an offshore movement of sand and shingle from south-east to north-west some several kilometres seaward of the west Wight coastline. This may be a source of partial replenishment of Pot Bank. The general conclusion that may be drawn is that Pot Bank has not been a source of supply to shingle beaches along the coast between Freshwater and Brook Bay. Thus, the impact of dredging may be provisionally discounted on the basis of circumstantial data from this (and adjacent) areas. This is supported by a closely reasoned case using data relating to tidal streams, wave refraction patterns and seabed levels (Geodata Unit, 1989). The same conclusion, based on work by Hydraulics Research (1977) on dredging of the Shingles Bank and Dolphin Bank, as well as Pot Bank, is reported in the section on Christchurch Bay.
South-East and South of Dunnose
Hydraulics Research (1977b, 1984, 1987) investigated the effects of proposed dredging for gravel at a site approximately 6km due east of Dunnose, at a mean water depth of 20m. It was concluded that gravel up to a median diameter of 25mm may, exceptionally, be mobilised in water depths of up to 22m, but is normally immobile below a depth of 15m. The prevailing wave climate would be incapable of moving gravel from that distance and at that depth onto the beaches of Sandown Bay. Based on wind fetch values, it was calculated that maximum wave heights vary from 4.63m (90 degrees) to 7.41m (240 degrees).
5. Summary
The entire southern coast of the Isle of Wight is subject to erosion, although its intensity varies due to wave exposure and its cliff morphology and behaviour varies according to the nature of the sequences of predominantly soft rock ground forming materials. Wave exposure is greatest along the southwest coast where there is a relatively narrow southwest facing fetch extending into the northeast Atlantic that coincides with the prevailing wave approach direction. By contrast, Sandown Bay in the east is sheltered from southwest approaching waves and is exposed primarily to moderate energy waves generated locally in the central and eastern English Channel. Sandown Bay and the South West Coast are composed of soft Wealden and Lower Cretaceous sediments, forming cliffs of moderate height, whereas the high "Undercliff" coast is characterised by sequences of soft Lower Cretaceous Sediments capped by permeable Greensand and Chalk lithologies within classic landslide generating sequences. The coastlines can thus be sub-divided into three distinct behavioural units summarised as follows:
Sandown Bay
1. The bay has formed through preferential erosion of soft, sandy sediments at its core and is occupied by sandy beaches and is anchored by headlands. An equilibrium planform does not appear to have been achieved yet by the bay so the tendency for erosion is likely to continue.
2. Cliff erosion constitutes the main source of sandy shoreline sediments, but the bay does not appear to be a location of long-term accumulation commensurate with the erosion yields.
3. Drift is northward within the bay, although there is no major sediment accumulation at Culver Cliff, suggesting that the bay contributes sediments elsewhere and is not an entirely closed system. Evidence suggests that sediments have been supplied to Ryde Sands, but is uncertain whether this link still operates via a nearshore transport pathway, or whether it was accomplished by shoreline drift operating at earlier stages prior to the emergence of Culver Cliff as a headland transport boundary.
4. Long term maintenance of the beaches of the bay is therefore critically dependent upon continuation of cliff erosion inputs.
5. Continued cliff retreat around Luccombe will cut further into the flanks of Shanklin and Luccombe downs and is likely to further re-activate relic landslides, potentially leading to rapid landward progressions of cliff top instability by several tens, or even hundreds of metres within specific events.
6. Without improved protection the Yaverland barrier beach could be susceptible to breaching. An extensive low-lying area of the E. Yar valley could become inundated to generate a tidal prism sufficiently large to maintain a new permanent tidal inlet.
The "Undercliff"
1. The entire coast is formed within a zone of massive relic landslides subject to marine erosion at their toes.
2. Marine erosion has removed considerable quantities of supporting material and actively eroding sea-cliffs have formed. Although the relict landslides have been relatively stable in historical times, there is increasing evidence that the likelihood of reactivations is increasing as support continues to be removed at the toes. Defences focussed upon Ventnor function directly to halt toe erosion and also to provide support to the toe of the coastal slope that is intended to reduce occurrences of instability within the relict landslides above.
3. Discrete landslide units of differing behaviour can be identified giving a short to medium term basis for prediction of ground behaviour. Some of the relict landslides are deep-seated, and may interlock with other relic slides further upslope such that stability may be mutually dependent and potentially large areas could become at risk following initially modest reactivations at the toe. Reactivations appear to be delayed where robust cliff toe protection in the form of engineered sea walls reduces recession of sea-cliffs. Conversely, reactivations are likely within the Undercliff above actively eroding sea-cliffs, especially at locations where the Undercliff is relatively narrow. The landslides are also sensitive to intensive or sustained rainfall due to its effects upon internal groundwater pore pressures and fluctuations of levels. The lag effect between cause and effect in this context is the subject of ongoing research.
4. The overall evolutionary morphological model would suggest eventual reactivation of all landslides, excavation of residual debris and then the initiation of new failures within in situ geological materials leading to renewed recession of the backscar. The likely timescale for such events is difficult to estimate (probably 100s of years), although it appears that processes operating towards full slope reactivation are occurring more rapidly in western than in eastern parts
5. Much debris derived from toe erosion does not stay at the cliff toe, but instead is rapidly removed seaward. As a consequence, the many cliff toes remain susceptible to further erosion even though residual sandstone boulders protect and armour some of the shoreline.
6. Small quantities of flints and chert gravel are released which, together with the sandstone boulders, remain on the shoreline and form small pocket beaches retained by minor headlands. Drift is weakly eastward along this frontage, but it is uncertain whether the pocket beaches are self-contained, or whether coarse sediments can periodically pass down drift form one to another. The latter could explain the widespread distribution of characteristic pea gravels on each pocket beach.
South West Coast
1. This coast comprises a long eroding cliff frontage that delivers large quantities of predominantly sand and clay sediment to the shoreline.
2. Beaches are composed of coarse sand and gravels, but are rarely especially high or well-developed, affording very limited protection to cliff toes. The majority of the material delivered is removed from the shoreline and its fate is uncertain. Sands may contribute to the nearshore bed and suspended sediments could be transported into the Solent, Poole Bay, or become moved greater distances along the Channel.
3. As the coast is eroded a widening dissipative shallow nearshore slope, and/or gently sloping shore platform is formed.
4. The high cliffs west and immediately east of Blackgang appear to becoming more unstable and accelerating in recession rate. They can be expected to become increasingly active in future, eventually leading to major new backscar failures. Many other soft rock cliffs along this coast are also likely to be susceptible to accelerating recession, especially in response to future climate change. Accelerated landsliding and cliff recession would considerably increase the delivery of sediments to the shoreline.
6. Opportunities for Calculation and Testing of Littoral Drift Volumes
Data collected by the Defra-funded National Network of Regional Coastal Monitoring Programmes is pivotal for future improvement in estimating beach change. The Programmes consist of topographic beach surveys, nearshore bathymetry, aerial photography, lidar, coastal hydrodynamics (waves and tides) and terrestrial habitat mapping. Specifications for data collection are consistent for all regional programmes and the data and analysis reports are made freely available under the Open Government Licence from www.channelcoast.org.
The Southeast Regional Coastal Monitoring Programme commenced in 2002. The Lead Authority is New Forest District Council, with data collection, analysis and reporting led by specialist teams at the Channel Coastal Observatory (CCO), Canterbury City Council and Adur and Worthing Councils. Although at present a 12 year time series of data has been collected, longer term Coastal Monitoring Programme data, when combined with other data sets, academic research and historical studies may enable sediment budgets, transport rates and directions to be identified and/or validated in the future.
The discontinuous nature of the shoreline of this unit with its headlands and pocket beaches in means that it is unsuited generally for definitive studies of drift. There are, however, opportunities to study drift occurring within Sandown Bay. An initial approach would be to develop a numerical model of littoral drift potential at a series of points along the full beach length based on an analysis of a long-term (greater than 20 years) hindcast wave climate.
Uncertainties encountered in applying numerical model studies would include:
I. Imprecision in the selection of synthetic wave climates. Shelter provided by the Dunnose headland and the tendency for significant wave energy to approach from south west and south east direction sectors introduces complexity in the waves actually experienced at the shore. These difficulties could be overcome by calibrating a sample of hindcast data against corresponding field measurements from the Sandown Bay wave buoy operated by the Channel Coastal Observatory [Sandown Bay Wave Buoy data 2004-12];
II. The problem of selecting a representative sediment gain size on the sandy and mixed gravel beaches (sediment mobility is highly sensitive to grain size);
III. The large differences likely between potential drift and the drift actually occurring that are controlled by sediment availability and interception by groynes. Beach volume changes provide good evidence of actual transport, but very little high quality profile data is available for this whole embayment.
7. Research and Monitoring Requirements
The Southeast Regional Coastal Monitoring Programme commenced in 2002. Analysis of the data between 2006 and 2012/13, combined with other datasets, academic research and historical studies has enabled sediment budgets, transport rates and directions to be identified or verified. However, at certain sites either due to a lack of long-term data, data coverage or sedimentological information (e.g. composition and proportion of beach grade material arising from cliff erosion), quantification of sediment transport rates of gravel and sand has not been possible.
Notwithstanding results from the Southeast Regional Coastal Monitoring Programme, and the summarised information collated in the Isle of Wight SMP2 (Haskoning, 2010), recommendations for future research and monitoring that might be required to inform management include:
1. To understand beach profile changes it is important to have knowledge of the beach sedimentology (gain size and sorting). Sediment size and sorting can alter significantly along this frontage due to variations in cliff supply, beach transport, sorting and beach management. Ideally, a one-off field-sampling programme is required to provide baseline quantitative information along the full shoreline, together with a provision for a more limited periodic re-sampling to determine longer-term variability. Examination of longshore and onshore-offshore grading of the various sediment parameters can be employed to indicate or confirm directions of transport, sources of sediment, including the possibility of supply from offshore to some Undercliff pocket beaches and possible residence (storage) timescales.
2. Shoreline conditions along the southern Undercliff are especially critical in determining the protection or exposure of the cliff toes that provide vital support for large areas of the landslide complex above. It is recommended that geomorphological mapping be undertaken of the sea cliffs, beaches and intertidal foreshore between Luccombe and Rocken End. It should seek to record the distribution of sheltering boulder aprons and highlight exposed areas. It should also attempt to identify the nature of sea cliff sediments and their likely erosion resistance and recession history (see recommendation no. 7 below). It would be beneficial to combine work with the foreshore sediment sampling recommended above and to integrate results with details of the nearshore zone derived from the hydrographic surveys planned by the Channel coast Observatory. An overall aim would be to compile a map of toe erosion potential that could complement and inform some of the studies already completed within the undercliff of ground movement potential.
3. Understanding of inputs of beach forming sediment from coast erosion, would be enhanced by cliff section mapping and sampling of deposits to reveal the detailed thickness, composition and variability of the lithological units (especially the Plateau and Valley Gravels, and flints and cherts within insitu Cretaceous strata - major local sources of beach gravel) occurring along the cliff tops. Sampling of such deposits could thus reveal particle size distributions, and be compared to similar analyses of stable beach materials. Quantitative information on cliff input should be coupled with details of beach sedimentology to assess the smallest size sediment grades stable on the beach and thereby determine the proportion of cliff input capable of contributing long term to beach volumes.
4. Littoral drift rates and volumes should be estimated in broad terms using details of cliff, and shoreface erosion inputs and beach volume changes. This information should then be integrated downdrift from the eroding cliff sediment sources to derive drift estimates based upon beach volume change. Studies could be undertaken for (i) Compton Bay, (ii) Brightstone Bay, (ii) Chale Bay (iii) the southern Undercliff coast and (v) Sandown Bay (see also Section 7). Analyses of volumetric data on cliff inputs, drift and beach storage are also required to begin to create a sediment budget framework for this coastline. Only by progressively improving the quantitative understanding of the system will it be possible to learn why beach levels have been diminishing along several of the protected frontages e.g. Sandown Bay. Using such information, remedial beach management should be able to be undertaken more confidently. In some cases, e.g. along the Undercliff coast, there appears to be little exchange of sediment between adjacent bays or coves, suggesting the possibility of a small-scale sequence of isolated compartments. In other cases e.g. Sandown Bay it could be that beaches are dependent upon supply from eroding cliffs. It is important to test these possibilities and assemble more definitive evidence concerning the operations of the coastal systems.
5. Sediment discharge from coastal valleys, or chines, may also make a contribution to the sediment budget of local beaches; its quantitative significance is unknown, but presumed to be comparatively insignificant. In the longer term, research might be focussed on the magnitude and frequency of large flows within chine systems.
MAP 24a (SW)MAP 24b (SE)LITERATURE REVIEWPHOTOS
