1. Introduction
Christchurch Bay comprises a shallow embayment (average depth of -7mCD with the 10m water depth contour between 3 to 5km offshore) bounded by Hengistbury Head (Photo 1) to the west and Hurst Spit (Photo 2) to the east. In comparison to Poole Bay, the shoreface is wider and shallower, promoting wave shoaling and refraction. The landward margin is pre-dominantly a currently or previously eroding cliffed coast of 5-35m in height e.g. (Photo 3), and includes Christchurch Harbour confined by low spit beaches near its western extremity (Photo 1). The study area extends several kilometres offshore to depths of between -6 and 14mCD to include the submerged rock platform of Christchurch Ledge and the sediment accumulations of Dolphin and Shingles Banks. The bay is of relatively recent origin, formed by coastline retreat during the mid to late Holocene transgression, although its contemporary geomorphology has been influenced by earlier geomorphological events.
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.
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). Two swath bathymetry surveys provide 100% coverage of Christchurch Bay. Through the Southeast Regional Coastal Monitoring Programme, the inshore 1 km zone of Christchurch Bay was completed in 2010. Further offshore, Poole and Christchurch Bays were surveyed in 2012 as part of the Maritime and Coastguard Agency’s UK Civil Hydrography Programme.
1.1 Coastal Evolution
Pleistocene evolution was dominated by the Solent River which flowed across the floors of Christchurch and Poole Bays and eastwards through the Solent (Everard, 1954; Dyer, 1975; Velegrakis et al. 1999). During climatically-controlled fluctuating sea-levels of this period, superimposed upon possible neotectonic uplift (Maddy, et al., 2000), the river and its tributaries deposited a sequence of progressively wider gravel-floored channels and terraces which mantled the Tertiary deposits of Christchurch Bay (Nicholls, 1987; Allen and Gibbard, 1993). A critical factor in the evolution of Christchurch Bay was the breaching of the Chalk ridge which previously extended between the Needles and Handfast Point, Purbeck. Opinion is divided as to the date of this event. Until recently, the prevailing view favoured breaching in the early Holocene (e.g. Everard, 1954). Recent studies of the rates of Chalk erosion; remnant buried and infilled palaeo- channels located in a west to east sequence offshore; the chronology of the lower terrace sequence of the Stour and Avon; early human occupation sites around Christchurch Harbour, and the depth, relief and inclination of the planation surface which truncates the Purbeck-Wight Chalk ridge indicate an early to mid-Devensian breach (West, 1980; Wright, 1982; Nicholls, 1985; Nicholls, 1987; Allen and Gibbard, 1993; Brampton et al. 1998; Velegrakis et al. 1999; Maddy, et al. 2000). An extension of the River Frome may have cut a gap through the western part of the ridge by the mid-Devensian, but later (Holocene) denudation of this feature appears to have utilised cols created by the headward erosion of rivers originally flowing southwards from its crest. It is suggested that much of Poole Bay was eroded during the mid to late Devensian, but Christchurch Bay was largely protected by the resistant strata (including ironstone concretionary seams) of a previously much more extensive Hengistbury Head. Velegrakis (1994) and Brampton et al., (1998) describe a 2m thick gravel ridge approximately 12km south of the Needles that both tentatively interpret as a relict barrier structure. This might define the position of the ancestral coastline in mid-Devensian times, when sea-level was between -40 and -60 m OD, prior to the final breakthrough of the chalk ridge. It was not until the early to mid-Holocene sea-level transgression (12-5000 years BP) that the Chalk ridge was removed, and the barrier between the two bays (now forming Christchurch Ledge) was cut back, and rapid erosion of Christchurch Bay proceeded (Wright, 1982; Bray and Hooke, 1998b). Excavation of Christchurch Bay was also facilitated by its connection with the Western Solent between 8400-6500 years BP, which created strong tidal current scour in the eastern part of the bay (Nicholls and Webber, 1987a). Simultaneously, wave and tidal current transport both out of, and towards, the West Solent entrance initiated the Shingles Bank; this has grown subsequently to its present volume of 40-60 million m³.
Both Christchurch and Poole Bays have a log spiral/zeta curve planform and concepts of crenulate bay formation have been applied, based on studies conducted in the United States (Wright, 1980, 1982). This analysis reveals that Christchurch Bay initially had an immature form which was an unstable configuration relative to a wave climate characterised by dominant waves from the south-west. An equilibrium plan shape has still to be achieved, hence continuing retreat. It is also apparent that Hengistbury Head has performed an important role in "anchoring" the planform of both Poole and Christchurch Bays, so geological controls are also important. (Halcrow, 1999; Brampton, et al., 1998). By comparison to Poole Bay, most of Christchurch Bay is more exposed to swell waves, and has a higher energy wave climate, despite the effects of wave refraction.
Cliff and shoreface erosion of Poole and Christchurch Bays, coupled with tidal scour of the western approaches to the Solent, released very substantial quantities of sediment, including gravel from river channels and valley terrace deposits as sea-level rose. It is thought that much of this material was transported eastward by littoral drift to form a succession of prototype forms of Hurst Spit (Nicholls and Webber, 1987a; Halcrow, 1999). The spit is probably not a classic multi-recurved form developed solely by littoral feed, but has evolved in a complex manner involving barrier transgression over a series of earlier Pleistocene gravel terraces and Holocene estuarine deposits, which themselves supplied sediment (King and McCullagh, 1971; Nicholls, 1985; Nicholls and Webber, 1987a; Bradbury, 1998). The spit is not a sediment sink because rapidly increasing depth and strong tidal currents at the distal end remove material. It therefore results from a complex interaction between sea level rise, sediment supply, storm overtopping events and the substratum (i.e. its capacity to support and contribute to the spit). Longitudinal extension has probably been controlled by offshore water depths and rates of sediment loss. Hurst Spit has probably receded relatively uniformly over the past 4000-5000 years since sea-level approached its present position (Nicholls and Webber, 1987a). Due to the complexity of controlling factors (and the fact that it is not a sink), it is particularly sensitive to change. Its response has been to vary its rate of recession, with periods of more rapid retreat being associated with phases of diminished sediment supply and high magnitude, low frequency storm surges. (Nicholls, 1985; Bradbury, 1998).
Two possible sinks are envisaged for coarse sediments released from Christchurch Bay, both involving net eastwards littoral drift to Hurst Spit and subsequent removal offshore. Some sediment may be transported from Hurst Spit to the West Solent (Dyer, 1971), but the major sediment flux appears to be south-west from Hurst Spit to feed the Shingles and Dolphin Banks and Dolphin Sand via the Needles Channel (Dyer, 1970; Nicholls, 1985; Velegrakis, 1994). The thickness of some of these accumulations has recently been determined by Velegrakis (1994) enabling approximate volumetric calculations. It is concluded that Christchurch Bay has been a virtually closed bedload transport system since the late Holocene except for probable (but diminishing) littoral drift input from Poole Bay and relatively small losses to both the outer bay and the West Solent (Nicholls, 1985). Recent budget changes have resulted from human activity, including beach and offshore aggregate mining, cliff stabilisation, reduction and control of longshore transport and beach replenishment. These have had both positive and negative effects on the quantities and rates of coastal sediment circulation, (as discussed in subsequent sections), but the system remains a quasi-isolated system, with minimal fresh input and relatively small losses in the context of the overall budget (Velegrakis, 1994). SANDFLOW model simulation confirms this (Brampton et al., 1998).
1.2 Hydrodynamic Regime
Wave and tidal current action are the dominant sediment transport mechanisms and the small tidal range serves to concentrate nearshore wave energy into a narrow zone.
(a) Wave Climate
The Southeast Regional Coastal Monitoring Programme measured nearshore waves using a Datawell Directional Waverider buoy deployed at Milford-on-Sea in 10mCD water depth, from 1996 to 2012. Prevailing wave direction is from SWbS. Average 10% significant wave height exceedance is 1.31m.
An offshore wave climate was compiled for the more energetic conditions in east Christchurch Bay using a hindcasting technique based on 15 years of Portland wind data and direct measurement from the BMT tower, offshore Milford-on-Sea (Hydraulics Research, 1989a and b). Prevailing direction was south-westerly and waves exceeding 1.0m were predicted for 31% of the time, and those exceeding 3.0m for 2.6% of the time. Extreme value analysis for waves of longest fetch (225-255EN) revealed offshore maximum significant wave height of 5.9m for a 1 year return period, and 7.2m for a 20 year return period. All waves exceeding 6.0m in height moved in from the south-west or west-south-west. HR Wallingford (1999a) using both actual and synthetic wave data obtained from the above, and subsequent, studies calculated that 28% of all waves approached from the south west, 13% from the south-south-west, and 11% from the south-east. Wave height and direction are modified inshore by shoaling and refraction; studies of these processes were conducted by Henderson (1979) using a five year wave-rider record (offshore Southbourne, in Poole Bay) for calibration. These studies and HR Wallingford (1999a) indicated high concentrations of wave energy on Hengistbury Head and at Barton. By comparison, the western sector of Christchurch Bay is less affected by south-westerly swell waves, but more exposed to waves approaching from the south-east. The complex offshore bathymetry of Christchurch Bay (particularly the shoals of Christchurch Ledge, Dolphin Bank and Shingles Bank) exerts a major refraction and focusing effect on incoming waves so that the resultant nearshore wave climate is spatially complex and difficult to model. Wave energy divergence occurs between Mudeford Spit and Highcliffe. (Henderson and Webber, 1979; Henderson, 1979). Gao and Collins (1994b) state that immediately offshore The Run, Christchurch Harbour, the most frequently occurring significant wave height is 0.6m, with 50% of recorded wave heights less than 0.25m. HR Wallingford (1999a) calculate that maximum wave height offshore The Run, for a one year recurrence, varies between 2.64m (155° approach) and 4.50m (215° approach). Halcrow (1999) state that the maximum inshore significant wave height varies between 4.1m (one year return frequency) and 5.3m (one in 50 years recurrence). Both inshore and offshore wave climate affecting the shoreline at Milford-on-Sea and along Hurst Spit have been determined by numerical modelling, calibrated by data obtained from the BMT wave rider tower for specific periods (Hydraulics Research, 1982, 1989a and b; Bradbury, 1998; Halcrow, 1999). For Milford-on-Sea and the proximal end of Hurst Spit, Hydraulics Research (1989b) calculate mean inshore significant wave height to be between 2.47m (1 year return frequency) and 3.76m (1 in 100 year recurrence). This was based on using the INRAY model to transform data from the tower site 1.5km offshore (in a mean water depth of 8m). In a later study, based on combined numerical and physical modelling, HR Wallingford (1993) concluded that mean inshore significant wave height, with a 1 in 10 year frequency, was 4.12m; Wimpey (1994) used a higher resolution version of INRAY to recalculate the same data set; this study proposed that a 1 in 100 year maximum wave height of 5.98m might be experienced. The equivalent value for offshore wave conditions, obtained using OUTRAY, was 6.15m (240° approach direction). Hague (1992) proposed a range of 6.5-7.00m for approaches between 210-270°. His work, which used previous data sets supplemented by subsequent records (Hydraulics Research, 1989b), proposed a maximum offshore significant wave height of 3.82m, with an annual frequency. This study revealed that the prevailing mean significant wave height outside the nearshore zone was between 0.5 and 1.00m (approaching from between 240 and 270°). A recent re-application of the HINDWAVE model in combination with TELURAY was calibrated using all available data sets on wave and water level conditions (HR Wallingford, 1999a). For mean offshore significant wave heights with a one year recurrence, the derived values were: 1.38m (155°); 3.60m (245°) and 4.15m (215°). Bradbury (1998) determined maximum 1 in 1 year maximum significant wave height to be 5.91m (135-240°). This value was derived from previous models, but was calibrated using improved digitised hydrographic data. For Milford-on-Sea, analysis using SANDS software, of 3 hourly offshore wave data between 1996 and 2010 (Channel Coastal Observatory, 2012) derived a maximum significant wave height of 3.47m for a one year return period, rising to 4.09m for 10 years and 4.68m for 100 years. Refer to Royal Haskoning (2011) for rose diagrams of wave directions of approach and significant wave heights recorded by the Waverider buoy offshore Milford, 2002-2010.
The comprehensive scheme of protection and stabilisation of Hurst Spit, completed in 1996 (see Section 4.3) generated specific studies of the variation of wave climate along its proximal to distal ends. These are set out, and analysed, in Bradbury (1998), who proposes a 1 in 100 year mean offshore significant wave height of 4.14m, approaching from 240°, for the proximal 800-1000m sector. For inshore waves, with an annual frequency, maximum significant wave height varies between 3.57m (240°) and 2.89m (210°). Halcrow (1999) used models previously employed by HR Wallingford (1989a; 1993) but applied them to UK Meteorological Office synthetic wave data. For the proximal sector of Hurst Spit, this study determined that the most frequently occurring nearshore wave heights are between 0.1 and 1.0m for all approach directions combined.
For the distal end of the spit (Hurst Point and Castle), Bradbury (1998) determines that the 1 in 100 year mean significant wave height (240°) is 3.10m. For inshore conditions, extreme wave heights with a probability of recurring once a year are 2.10m (210°) and 2.68m (240°). For the 240° approach direction, Halcrow (1999) deduced mean wave heights to vary from 1.6m (1 in 1 year) to 2.60m (1 in 50 years). Some 75% of all waves arriving at the distal end, from all directions, have a mean height of 0.1 to 1.5m.
It is therefore apparent that there is a progressive north-west to south-east reduction in nearshore wave energy along the main "corridor" of Hurst Spit. Taking into account all previous studies, Bradbury (1998) suggests that mean maximum nearshore wave height declines from 4.1m to 3.1m. The main reason for this is the attenuating/dissipating influence of Shingles Bank and North Bank, setting up complex refraction and wave train "crossover". Wave shoaling and breaking (at low water states) induced by the complex bathymetry of the banks and channels seawards of the distal sector reduces the height of offshore waves by almost precisely one third (Bradbury, 1998). It is probable that nearshore wave conditions have varied as the offshore banks, shoals and channels have evolved. Contemporary monitoring (New Forest District Council, 1992; 1997-2001) reveals constant fluctuations in bank morphology, but specific impacts on wave climate are probably fairly limited over the medium timescale (1-10 years). More important is the impact of high magnitude, low frequency storm and tidal surges on water levels, and therefore water depths, on shoaling and refraction. This is apparent from analysis of the 1 in 100 year recurrence storms of October and December 1989 (Bradbury, 1998). The latter generated a higher water level, but prevailing inshore wave heights (2.90 m at the neck of Hurst Spit) were less than two months previously (3.60 m) because of reduced wave height peaking. Monitoring of the frequency of maximum nearshore wave heights since 1996 (New Forest District Council, 1997-2001) will allow further analysis of this factor.
For the Bay as a whole, dominant waves approaching from the west to southwest are therefore subject to spatially and temporally variable refraction with the result that nearshore wave conditions vary both along the coast and with time. Halcrow (1999), using simulated wave data derived from the UK Meteorological Office model and local bathymetric data concluded that wave climate at all points along the shoreline was principally a function of wave approach direction. In a series of papers (Bradbury et al., 2006, 2007, 2009 and 2011) analysis of wave spectra recorded by a Waverider buoy offshore Milford-on-Sea revealed that bi-modal wave conditions (Atlantic swell and locally propagated wind waves) were incident for up to 20% of defined periods, and were characteristic of more than 40% of winter storms. Refer to the later section (5.3.7) on Hurst Spit beach for further discussion of beach response to bi-modal events.
(b) Tidal Regime
The tidal range of Christchurch Bay is the lowest along the south-central Channel coast. At the entrance to Christchurch Harbour, the mean spring range is 1.4m, reducing to 0.8m during neap cycles, although there is some attenuation of wave form as it moves through the narrow channel connecting the harbour with The Run immediately seawards.
Tidal currents are extremely rapid at Hurst Narrows (up to a maximum 3m per second at the surface and 2.5m per second close to the seabed) and are capable of significant transport of coarse sediment (see also unit on West Solent and Webber (1980a)). Currents diminish offshore and longshore so that dominant south-eastward currents on the southwest face of Hurst Spit attain peak velocities of 0.8-0.9m per second (Nicholls, 1985). Metering surveys show that currents generally diminish to less than 1m per second to the west of Shingles Bank (Velegrakis, 1994) and are below 0.5m per second in the central area of the Bay (Brampton et al., 1998). In the vicinity of Highcliffe and Barton, dominant ebb tidal currents flow eastward at up to 0.3-0.4m per second (Lacey, 1985) to give a weak near-seabed clockwise tidal gyre, defined mathematically from tidal flow residuals, for the Bay as a whole (Riley et al., 1994; Riley, 1995; Halcrow, 1999). Rapid tidal currents are thus restricted to the eastern extremity of the bay, except for Christchurch Harbour entrance, where ebb flow coupled with fluvial discharge can create maximum velocity currents approaching 2.5m per second (Tosswell, 1978; Gao and Collins, 1994a and b, 1995c; HR Wallingford, 1999b). Here, the duration of the flood tide is 6.4 hours, whilst the ebb phase occupies 6.01 hours. Gao and Collins (1994b) determined that the threshold velocity for the entrainment of fine sand on the bed of the harbour entrance channel is between 1.7 and 1.9m per second. Mean tidal range within the harbour is lower than on the adjacent open coast because of attenuation effects set up by The Run (Hydraulics Research, 1987). In both Christchurch Bay and Harbour, a "double high water" effect (i.e. a high water long stand, followed by a second high water peak) is discernible, though less well pronounced than in Poole Bay. This contributes to the ebb-dominant asymmetry of the tidal regime of Christchurch Harbour.
2. Sediment Inputs
2.1 Marine Inputs
» EO2 · F1
As Christchurch Bay is widely regarded as a largely self-contained sediment circulation system (Nicholls, 1985; Bray and Hooke, 1998b; Halcrow, 1999; Royal Haskoning, 2011), it is difficult to envisage any significant externally-derived marine input. This view is supported by a variety of evidence:
- Chart comparisons of the Pot Bank dredging area south of the Needles covering the period 1937-72 revealed seabed lowering of 1m (Hydraulics Research, 1977). The surveyed seabed change was only 18% of that removed by dredging during this time (4 million m³), thus providing indirect evidence that material was leaving the area. Despite this net loss, analysis suggests that Pot Bank is a sediment sink receiving sediment from the Needles Channel and Shingles Bank, with no return transport (Hydraulics Research, 1977; Brampton et al., 1998). This conclusion is partly based on modelling of tidal residuals, and near sea-bed current velocities operating below gravel mobilisation thresholds.
- Offshore survey involving echo sounding, oblique asdic sonar and grab sampling have revealed an abrupt transition in seabed morphology and materials south of Dolphin Bank and Sand (Dyer, 1970; Velegrakis, 1994). Both banks are characterised by predominantly sandy glauconite-rich sediments with bedforms indicating a net overall westward transport. Further offshore, in deeper water, gravel bedform features indicate southward transport, and limonite-rich sediments suggest possible feed from the nearshore/offshore zone of southwest Wight. Sediment transport between the two areas appears extremely limited, and thus output from Christchurch Bay appears much more likely than input.
- Christchurch Ledge is widely recognised as a barrier to net onshore bedload sediment transport into the western part of Christchurch Bay from further west and south-west (Dyer, 1970; Wright, 1982; Brampton et al., 1998), although evidence is not conclusive.
- Chart comparisons covering the whole of Christchurch Bay for the period 1880-1968 revealed overall loss of 505,000m³ per year of sediment (Lacey, 1985). Beach profiles and nearshore hydrographic survey covering the period 1974-90 also indicated loss of sediments (Webber, 1980b; Halcrow, 1980; Lacey, 1985; Velegrakis, 1994). This information again suggested net sediment output rather than input. Although the accuracy of hydrographic surveying was frequently insufficient to determine changes over wide areas for short time periods, the long-term bed erosion identified by Lacey (1985) is equivalent to uniform lowering of 0.39m, which was probably within the limits of survey accuracy.
- Analysis of suspended sediments in Christchurch Bay using Landsat imagery revealed a marked seasonal trend corresponding to peak cliff erosion input during winter storms (Lacey, 1985). The majority of suspended sediments are probably derived from this source and no evidence of any independent/supplementary marine input is available. However, research has been insufficient to draw any firm conclusions.
Potential marine input can nonetheless be identified, namely:
Ebb tidal currents are of shorter duration and more rapid at Hurst Narrows (Webber, 1980a), thus a dominant southwest transport pathway extends along the Needles Channel (Dyer, 1970; Nicholls, 1985; Velegrakis, 1994). Previous research suggests net eastward transport into the West Solent, but this does not fit with the energetics of the tidal regime. Thus it is postulated that tidal current-moved sediment output is probable from the West Solent, thereby feeding the vigorous southwest transport path at the extreme eastern margin of Christchurch Bay. (See Section 6 for a full account based on bathymetric, sedimentological and morphological data).
The 2004 SCOPAC map had an F1 arrow in the vicinity of the Long Groyne, to represent wave driven offshore to onshore sediment transport. This was attributed to onshore feed of sand and gravel between 476,000m³ per year (Lacey, 1985) and 559,000m³ per year (Henderson, 1979). These estimates are of low reliability as they are based solely on theoretical calculations of longshore wave energy flux with no consideration of site-specific littoral processes or sediment availability. Serious problems were encountered with wave refraction and energy diffusion in the area of complex bathymetry set up by Christchurch Ledge, and sediment transport equations have been calibrated with extremely limited data. No further evidence supports this F1 arrow in this location and it has therefore been removed.
The new 2012 updated F1 arrow is now located in the vicinity of the Mudeford Sandbank, a dynamic feature supplied by sediments from the nearshore area of Christchurch Harbour ebb tidal delta. The periodic breaching of this sandbank often supplies the eastern bank of Mudeford to Avon beach, which periodically extends to the west from wave driven offshore to onshore sediment transport. No volume is given as this transport mechanism appears to be episodic. Several studies have been conducted in Christchurch Bay using Woodhead Seabed Drifters (Clark and Small, 1967; Watson, 1975; Tyhurst, 1976; Turner, 1990); a total of nine experiments were undertaken between 1966 and 1975. These clearly revealed that Mudeford-Highcliffe was a major receptive area for drifters, indicating a potential for onshore transport to this coastal segment. Recoveries were generally between 38% and 75%, but as they relied upon the general public there were problems with the quantity and quality of information. This was only available for time and location of injection and recovery, so the intervening pathways, as well as the mechanisms of transport, are conjectural. It is uncertain exactly what drifters measure, because it has been shown that they are not necessarily reliable indicators of movement at or close to the seabed (Collins and Barrie, 1979). The experiments probably indicated onshore currents at the seabed, but the significance of this to sediment transport is uncertain; thus this pathway is of low reliability. Limited support is provided by study of the historical behaviour of Mudeford Spit using map and chart comparisons (Robinson, 1955; HR Wallingford, 1999) from which it is tentatively concluded that it has been sustained by onshore rather than longshore supply. (See Section 5.2 for detailed discussion).
One study using seabed drifters indicated a potential for onshore sediment supply at Milford and Hurst Spit (Clark and Small, 1967), but subsequent drifter experiments indicated an offshore southeasterly trend in this area (Watson, 1975; Turner, 1990). These studies are therefore contradictory and inconclusive; an assessment by Dobbie and Partners (1984) and Nicholls (1985) concluded that significant net onshore feed to Milford-Hurst was unlikely, particularly of gravel. Offshore survey by side-scan sonar, echo sounding and sediment sampling revealed asymmetric gravel ripples which indicated a net onshore transport vector over the western part of Shingles Bank (Velegrakis, 1994). This pathway supplies gravel to a dominant north-westwards sediment flow in North Channel (Dyer, 1970; Velegrakis, 1994), but the ultimate sink is uncertain. Coarse sediments are therefore highly mobile offshore Milford-Hurst, although an onshore transport supply to North Channel has been recognised (Velegrakis, 1994) which may not be a long-term feature. This information was represented in the 2004 SCOPAC maps as a F3 arrow, but has been removed in the 2012 update due to the contradictory and inconclusive information.
2.2 Fluvial Input
FL1 The Rivers Stour and Avon
Combined discharge of the Stour and Avon to Christchurch Harbour averages 30m³ per second with a minimum flow of 7.5m³ per second and a maximum of 220m³ per second (Tosswell, 1978; Gao, 1993; Gao and Collins, 1997a). ABPMer (2009) report 1 in 100 year peak flows on the Avon to be 127m³ per second, but up to 380m³ per second on the Stour. It is reported that suspended sediment load is relatively low due to catchment supply from Chalk aquifers and interruptions to flow, e.g. weirs (Murray, 1966; Tosswell, 1978; Rendel Geotechnics and University of Portsmouth, 1996), but is periodically supplemented by navigable river channel dredging to alleviate flood risk. In recent years, dredge spoil has been deposited elsewhere in the harbour area. Sampling surveys revealed sand and gravelly sandbanks adjoining the river channels in close proximity to their points of discharge into the harbour, and these may provide some supply (Gao, 1993). Annual siltation is reported in navigable channels in the harbour and accelerated rates during the winters of 1990-91 and 2003-4 on Grimsbury Marsh were alleged to result from dredging of the lower Stour. Mean fluvial discharge is approximately equal to the tidal prism, which indicates clear potential for sediment supply (Tosswell, 1978). Rendel Geotechnics and the University of Portsmouth (1996) calculate a maximum combined bedload input of some 150 tonnes per year, but actual quantity is probably less than 10% of this total due to numerous weirs and sluices on both rivers. Gao and Collins (1995b) calculate bedload input as 320-640m³ per year, based on a largely theoretical calculation. Suspended sediment discharge potential is close to 70,000 tonnes per year, but actual delivery is unlikely to exceed 10,000 tonnes per year. It is uncertain how much of this material is transported through the harbour entrance; some is understood to be retained within the estuary system especially in the northeast, where it contributes to inner harbour mudflat and grazing marsh edge sedimentation (Gao and Collins, 1995a).
There is no sediment delivery via Chewton Bunny, as it discharges via a culvert. Other streams to the east, notably Becton Bunny, contribute only very small quantities of predominantly fine sediments, despite their well-incised valleys and erodible catchments. It is suggested that fluvial inputs come mostly from the harbour dredging and beach replenishment activities and are not considered a natural source of sediment (Royal Haskoning, 2011).
2.3 Coast Erosion
» E1 · E2 · E3 · E4 · E5 · E6
Introduction
Much of Christchurch Bay is backed by rapidly eroding cliffs of up to 33m height and 120m between crest and toe, which provide a major sediment input where they are not stabilised and protected. Local rates of erosion are determined by geology, the longshore distribution of wave energy, the degree of cliff/coastal slope protection afforded by the natural beach or protection structures and cliff stabilisation measures. The cliff-forming strata comprise overconsolidated Tertiary sands and clays which dip 0.5 to 1° towards the east-north-east and strike nearly parallel to the coastline, so that progressively younger beds outcrop from Hengistbury Head (Hengistbury Formation) to Hordle and Milford (Headon Formation). There is also a slight inland stratal dip. These substrate materials are overlain by a mantle of Pleistocene Plateau Gravels and thin Holocene brickearth deposits. The nature of sediments supplied is therefore controlled by cliff lithology and it is possible that this has altered throughout the Holocene as coastal recession has truncated lithologically variable deposits. The Pleistocene gravels are particularly variable in composition, and it is suggested that previous cliff sections included deep infilled channels (Nicholls, 1985), contributing a gravel supply that was at least twice the volume that prevailed after coast protection. Lithological variations of 'solid' substrate also affect the rate of erosion where sandstones (permeable) overlie clays (impermeable) with critical groundwater pore pressures facilitating mass movements and slope failure (Barton, 1973; Barton and Coles, 1984; Rendel Geotechnics, 1991, 1998; High-Point Rendel, 2002; Barton and Garvey, 2011). Nicholls (1985) has proposed that over the past 2-3000 years, up to the introduction of protection and stabilisation measures, the supply of sand from cliff erosion was eight to ten times the volume that is received today. Protection measures and longitudinal structures such as outfalls at Becton have created several terminal groyne (outflanking) effects, accelerating downdrift erosion as a result of reduced supply to beaches protecting cliff toes (Jezzard, 2004; Brown, 2008; Brown and Barton, 2007). Lacey (1985) calculated that between 48% and 54% of all material derived from cliff erosion is fine grained, and is rapidly transferred permanently offshore via suspended transport.
Long-term recession has created a shore platform that widens progressively eastwards in response to exposure to wave energy. Posford Duvivier and the British Geological Survey (1998) provide some largely hypothetical figures for contemporary annual shoreface erosion (both vertical erosion and volume of yield).
Map comparisons covering the period 1896-1976 have revealed cliff top erosion at 0.42-0.77m per year to the north-east of the Long Groyne due to recurring slips (Photo 4). The higher rate was located almost adjacent to this structure and attributed to wave energy focusing caused by it (Hydraulics Research, 1986a). An armoured revetment and four rock groynes were constructed in the late 1980s and now protect the cliff toe, although input is still possible due to erosion by high energy storm waves approaching from the east/south-east. Small quantities of sediment are yielded from these cliffs, which comprise sandy clays capped by thin valley gravels. Artificial removal of talus is undertaken to maintain exposure of the internationally important stratigraphy and palaeontology of this GCR/SSSI site (Bray et al., 1996; Bray and Hooke, 1998).
The Hengistbury Head Management Plan (Bournemouth Borough Council, 2005) suggests that the rate of cliff erosion in the north-east of the Long Groyne is generally very low, with some areas of gullying where the cliff face has retreated in excess of 10m. Furthermore, it is suggested that erosion by pedestrians combined with wind and rain has caused lowering of the cliff top by at least 1m. The rate of cliff retreat is thought to be linked to large scale slips within the clay and silt zones of the Upper Hengistbury Beds (Bournemouth Borough Council, 2005). Regional Monitoring lidar difference plots (2005-2012) show notable erosion of the cliff face to the north-east of the Long Groyne; it is not likely that this sediment directly supplies the beach, but instead accretes as talus at the base of the cliffs.
The cliffs here are composed of Highcliffe Sands to the west (Photo 3), with an increasing thickness of underlying impermeable, overconsolidated and fissured Barton Clay to the east. The latter contains lenses of sand, which supply groundwater to the cliff face and thus facilitate landsliding. The succession is overlain by Plateau Gravels varying from 1.5 to 7.6m in thickness, creating strong seepage and gullying at its junction with the underlying sandstones.
The cliffs at Highcliffe have a long history of instability and retreat. Historical records indicate that retreat may once have been extremely rapid, and a possibly unreliable rate of 6m per year is quoted for the period 1760-1830 (Nicholls, 1985; Bray et al., 1996). This is supported by an account suggesting loss of up to 1km (10m per year) between the late 1700s and late 1800s (Mockridge, 1983). Using Ordnance Survey maps of the Highcliffe and Barton-on-Sea frontage between 1871 to 1960, cliff recession was estimated to be 1m per year (Halcrow, 1978b) this is further supported by an estimated rate of 1.05m per year for Highcliffe in the Poole and Christchurch Bays Shoreline Management Plan (Royal Haskoning, 2011). Subsequently, the cliffs became significantly more stable, probably due to improved drainage and toe protection by an accreting beach fed by the extension of Mudeford Spit (Mockridge, 1983). Mean recession of 0.18m per year was recorded over the period 1908-59 by comparison of maps and a cliff-top survey in 1958 (Wise, 1959). Rates were spatially and temporally variable with maximum recession for a short central cliff section of 0.95m per year for 1931-39 (Wise, 1959). The cliff toe periodically advanced and retreated over this period as basal landslide debris intermittently surged seaward and was then eroded (Barton, 1973). This phase of relative stability ended in the 1950s following the retreat and breakdown of the formerly extended form of Mudeford Spit, and thereafter the Highcliffe cliffline was characterised by increased frequency of landsliding and cliff-top retreat rates which increased to 0.68m per year for the period 1965-1975 (Barton, 1973; Mockridge, 1983; Tyhurst, 1985a, 1986).
The unmodified cliffs formed a bench and scarp profile with failure of the upper scarp by relatively deep-seated slumps and degradation, and transport of debris over the undercliff benches by mudslides (Barton, 1973; Mockridge, 1983). These processes supplied significant quantities of sediment to the foreshore between 1932 and 1968; Lacey (1985) calculated a supply of 4,400m³ per year of material greater than 0.08mm diameter, regarded as being stable on the beach and lower foreshore. This supply has now largely ceased due to the effectiveness of coast protection and cliff stabilisation schemes commencing in the late 1960s. These include construction of a revetment (usually buried beneath beach gravel), groynes, cliff drainage and slope regrading in 1973/74 and 1978/79; vegetation establishment over the undercliff to enhance slope stability, and combined beach nourishment and rock revetment and groyne construction in 1985 and 1991 (Photo 5). Although the cliffs between Friar's Cliff and Highcliffe Castle are unprotected, from 1986 onward cliff drainage was inserted further west towards Highcliffe Castle (Mockridge, 1983; Tyhurst, 1986; Christchurch Borough Council, 1991; Halcrow, 1999). Cliff input is not possible whilst these measures are effective, but the talus store in the vicinity of Friar's Cliff is potentially available as an input source.
Regional Monitoring lidar difference plots (2008-2013) show that short sections of exposed cliff are liable to erosion, but that this material may not directly supply the beach. The majority of clifftop along this frontage are vegetated, making analysis of lidar and aerial photography difficult.
This segment comprises naturally and rapidly eroding cliffs, in contrast to the now protected and stabilised adjoining cliffs at Highcliffe and Barton (Photo 6). It thus provides one of the few remaining natural input sources of beach sediment in Christchurch Bay (Mackintosh and Rainbow, 1996; Moore et al., 2003). The low gradient cliffs (shallow angle of repose) are composed of eastward dipping Barton Sand and Clay overlain by 1.5-3.0m of Plateau Gravel (Barton and Coles, 1984). It is reported that former exposures of these gravels included Pleistocene buried channel deposits up to 16-18m thick, but diminished as cliff retreat cut further landward (Nicholls, 1985). Lithological variations within the Barton Clay cause development of three preferred bedding plane shear surfaces marked by benches in the cliff profile. The rear scarp fails by deep-seated rotational slides extending to the uppermost shear surface and periodically down to the intermediate surface. The cliff top therefore continues to recede rapidly via recession heads and "breakaways", with single events removing up to 2m. The benches are sites of degradation and downslope sediment transfer due to small scale slips, bench sliding, slumping and mudsliding (Barton and Coles, 1984). Bench sliding is considered the dominant process, possibly accounting for over 90% of slope movement (Halcrow, 1999). Groundwater seepage at lithological junctions is a prime factor facilitating superficial flows and slides.
A long-term retreat of approximately 1m per year has been characteristic of these cliffs according to map comparisons over the period 1867-1959 (May, 1966; Nicholls, 1987; Hooke and Riley, 1987; Halcrow, 1978b). This conceals spatial variations of short-term rates of up to 5m per year. West (2013) suggests locally higher rates in the Naish Farm area because of the effect of sea defences to the west. Retreat over the period 1869-1939 was 0.3m per year, increasing to 0.4m per year by the late 1950s, 1.3m per year (1960s), 1.9m per year (1970s) and 2.4m per year (1990s) (Mackintosh and Rainbow, 1996). Following construction of groynes updrift at Highcliffe, this sector retreated by 60m or 3m per year 1970-1985 (Halcrow, 1999). This accelerated retreat is attributed to sediment starvation due to outflanking set up by the terminal groyne effect at Chewton Bunny (Bray, et al., 1996; Mackintosh and Rainbow, 1996). Sparshott (2001) derived a mean rate of cliff top recession for Naish cliffs of 1.12m per year, 1976-1998. Maximum and minimum rates of 1.55m per year and 0.45m per year, respectively, were determined for different sectors. Moore et al. (2003) calculated mean and maximum annual recession rates for unprotected cliffs at Becton Bunny and Naish Farm as 1.13m per year and 1.63m per year respectively, between 1940 and 2001. Mean and maximum rates were also calculated for protected cliffs along the same frontages 0.4m per year and 0.7m per year respectively. Cliff toe recession was calculated to be 0.65m per year over the same period, with a maximum of 1.22m per year recording the rapid erosion of a debris fan. This research, which was based on contour plotting from DTMs based on three sets of vertical aerial photography obtained over a 30 year period, revealed a reduction in the gradient of the coastal slope. It also demonstrated that there is close adjacency of relatively stable and unstable sectors of the cliff face, but that the zone of maximum morphological change moved progressively eastwards between 1976 and 1991. Together with other studies (Rendel Geotechnics, 1991; 1998; Mackintosh and Rainbow, 1996; High-Point Rendel, 2002; Moore et al., 2003), active slope degradation was identified as a largely winter season activity.
Integration of information on sediment composition exposed in these cliff sections with the above quoted mean retreat rates have enabled sediment yield to be calculated. Total supply of sediments >0.08mm diameter was 9,000m³ per year, in 1984 (Lacey, 1985). Sampling of the Plateau Gravel deposits revealed that 46% was gravel >8mm diameter (Indoe, 1984). Combining this with a mean deposit thickness of 2m, and a prevailing retreat rate of 2m per year (Barton and Coles, 1984) over the 1,300m eroding length of cliffline gives an approximate gravel supply of 2,400m³ per year, this volume is used to support the 1-3,000m³ per year E3 arrows, a reduction from the 2004 volume of >20,000m³ per year. It is thought that this higher volume estimate represented the total yield of predominantly fine grained sediment from cliff erosion is estimated at 58,000m³ per year (Posford Duvivier and British Geological Survey, 1998). Using SURFER, Sparshott (2001) calculated volume loss, below 30mOD, to be a mean of 179,000m³ per year, a figure which includes clay, sand and gravel. This applies to the entire sector between Chewton Bunny and the western end of the stabilised cliffs at Barton. Barton and Coles (1984) proposed that a total annual yield of 5,740m³ derived from a 200m segment of this cliffline in the early 1980s.
Recession of this cliffline between two protected frontages has created an embayment with an immature log spiral (crenulate) plan shape. Crenulate bay concepts can be applied to model the evolution of such features and indicate that a further 140m (Lacey, 1985; Webber, 1980b; Halcrow, 1980) to 240 m (Halcrow, 1999; Royal Haskoning, 2011) of erosion is required at the western extremity before a stable plan shape is established.
Analysis of Regional Monitoring lidar data (2008-2013) shows major erosion of the Naish frontage. The cliff top recession can be assessed using aerial photography with horizontal recession up to 15m (2005-2013). It is clear that this material directly supplies the beach, although the fine fraction is typically not retained by the beach and is lost as suspended material.
Historically, this segment was also extremely unstable and subject to rapid cliff top recession, measured at 0.75m per year for the period 1867-1959 (Hooke and Riley, 1987). Phillips (1972) calculated mean rates of 1.1m per year (1868-1907); 0.9m per year (1908-1936) and 0.5m per year (1937-1958) with rates as high as 3.0m per year for individual years. The Poole and Christchurch Bays Shoreline Management Plan (Royal Haskoning, 2011) estimated longer term erosion rates for Barton-on-Sea frontage of 1.05m per year and for Becton Bunny frontage of 1.30m per year. Moore et al. (2003) estimate a retreat rate of 1.03m per year for this sector of cliff line as a whole, based on analysis of multiple air photo sorties dating from 1940 to 2001. Singular events produced cliff recession of 2 to 5m approximately once every 5 to 13 years. Picksley (2012) contains detailed plots of historic positions of the cliff top edge on aerial photo cover for 1957, 1966, 1989, 2001, 2005 and 2008. Cliff form and processes were generally similar to those now exhibited in the unprotected segment to the west except for an important lithological distinction. Eastward dip of the stratigraphical succession gradually brings the impermeable Barton Clay down towards beach level at Becton Bunny, thereby increasing the extent of cliff outcrop of the overlying permeable Barton Sands and Plateau Gravels. Groundwater is directed towards the cliff-face at the interface between the sandy and clayey strata separating (i) the Upper and Middle Barton Beds and (ii) the Plateau Gravel and Upper Barton Sand including the Charma bed, promoting seepage and gullying (Daley and Balson, 2000; High-Point Rendel, 2002). The Plateau Gravels and Upper Barton Beds together form a partly confined aquifer, as does the Highcliffe Sands within the Middle Barton Beds). The sandy materials above this interface tend to fail by deep-seated rotational slides extending down to this shear surface/zone of weakness. (Barton, 1973; 1998; Barton and Coles, 1984; Fort et al., 2000; High-Point Rendel, 2002; Barton and Garvey, 2011). This surface formed a bench or undercliff characterised by degradation and seaward transport of sediment towards a steep sea-cliff cut in the underlying Barton Clay (Clark, Ricketts and Small, 1976). Increased cliff top recession of 2.0-2.4m per year for 1959-66 (Barton, 1973), despite groyne construction between 1939 and 1954 and a temporary increase in beach width due to the longshore dispersal of sediments composing Mudeford Spit after it’s breaching in 1935, endangered cliff top properties. This led to a more comprehensive coast protection scheme which was completed between 1966 and 1968 (Barton and Garvey, 2011) (Photo 7 and Photo 8). This involved deep drainage, slope re-modelling and regrading, and toe protection comprising groynes and a flexible timber revetment with retaining rock armour (Phillips, 1972, 1974; Wright, 1992). Cliff stabilisation over a frontage of 1.75km included steel sheet piling driven through the undercliff to intercept groundwater flow along the principal clay/sand interface, and seaward drainage of groundwater thus re-directed (Phillips, 1974; Clark, Ricketts and Small, 1976; Fleming and Summers, 1986; Wright, in Bray and Hooke, 1998). These measures were only partially effective because deep-seated failures in the undercliff in the winter of 1974/75 penetrated beneath the sheet piling causing its rotation, bowing and splitting (Clark, Ricketts and Small, 1976). It is probable these failures were along previously undetected slip planes within the Barton Clay. This would be similar to the landslide process west of Barton where movements are frequent on an upper slip plane, but intermittent on a lower shear surface (Barton and Coles, 1984). The result was two-fold; first, mudslides surged over the cliff and revetment and supplied sediment to the beach; second, cliff-top recession averaged 5m per year over a 150m front in 1975 (Clark, Ricketts and Small, 1976; Indoe, 1984; Wright, 1992; SCOPAC, 2011). Various emergency and long term remedial slope regrading and drainage works and beach stabilisation using five 400m spaced rock groynes/strongpoints (Photo 8) linked by a rock revetment, combined with periodic beach nourishment have been partially successful in preventing further major slides (Wright, 1992; Bray, et al., 1996), although subsequent detailed geomorphological mapping and site monitoring has revealed continuing and reactivated ground and sub-surface movements due to compound failures at lithological boundaries within the Barton Clay (evidenced by tension cracks and other diagnostic features), and shallow translational mudslides, e.g. at the Cliff House hotel and Barton Court in 1987, 1993 ,1996 and 2001 (Rendel Geotechnics, 1998; High-Point Rendel, 2002; Fort and Clark, 2002; Moore et al., 2003). Barton et al., (2006) and Barton and Garvey (2011) identify that the most dominant of the preferred shearing surfaces generating compound landslips is located near the base of the clay mineral rich D zone of the Barton Clay. They conclude that displacement results from lateral rebound response to progressive slope recession induced by coastal erosion (Rendel Geotechnics, 1998). Investigation has established that rainfall exceeding 80mm is sufficient to initiate movement, usually during the winter (Fort et al., 2000; High-Point Rendel, 2002). A major multi-plane slip at the western end of this sector in 1993 displaced the toe revetment 8m seawards and required the construction of new drainage and a major rock revetment to maintain slope stability by toe weighting (High-Point Rendel, 2002; Moore et al., 2003; Barton and Garvey, 2011) (Photo 9). Subsequent ground movements and displacements of previous remedial and protective structures have required slope regrading and improved drainage, with further measures anticipated by ongoing monitoring. The failure of the sheet piling in this central area and associated drainage resulted in the loss of the access road at the end of Sea Road. Barton and Garvey (2011) state that by the middle of the first decade of the twenty-first century almost 50% of the total length of the undercliff stabilised in the mid-1970s by the insertion of a sheet piled filter drain had been rendered ineffective by landslide reactivation. This, they argue, may be due to insufficient evaluation of the role of degradational features such as re-entrant mudslide gulleys and failure to identify the relative importance of specific shearing surfaces (such as the D zone at a shallow depth below the beach at the western end of this cliff complex). In general terms, therefore, whilst cliff foot erosion has been curtailed by protection structures, cliff face and cliff top sub-aerial weathering and structurally induced erosion continue as active agents of slope recession. Repairs to the culverted spring pipe beneath the Cliff House Hotel were required in 2012.
The eastern part of the cliff line towards Becton Bunny remains largely unprotected (Photo 10) and retreated in consequence of unconstrained toe erosion at 1.63m per year from the late 1970s to early 1980s (Lacey, 1985). Supply capacity is high because Plateau Gravel deposits attain a thickness of 5-9m (Nicholls, 1985) and the thick Barton Sand supplies material capable of contributing to the lower foreshore. Total supply >0.08mm diameter was calculated as 13,000m³ for 1984 (Lacey, 1985), although prior to the 1966-8 protection scheme it was much greater. The E4 erosion input volumes were >20,000m³ per year in the 2004 STS. This majority of this frontage is protected at the toe by rock revetments, which effectively block the direct supply of cliff material to the beach. There are two potential input locations where, after significant ground movement, fine clay material flows over the rock revetment onto the narrow beach. This is evident beneath Marine Drive West car park and the Cliff House Hotel. The volumes are unquantified, with episodic occurrence and likely to be of fine grained material lost in suspension. As a result the E4 erosion input volumes presented in the 2012 update are ‘no quantitative data’.
The role of groundwater in causing instability of the cliff and upper beach at Barton cliffs has long been appreciated. Results from ground investigations and monitoring, notably during the exceptionally wet winter of 2000/2001, show a clear relationship between ground water movement and wet weather periods (High-Point Rendel, 2002). Direct measurements of rainfall provides adequate information to predict the onset of movements (Fort, et al., 2000). Bradbury (2008) found a clear relationship between the onset of ground movements and 2 month antecedent rainfall. Once the threshold has been reached, which occurs at about 65mm, ground movements tend to continue until after the average rainfall levels reduce to about 40mm. Substantial movements over all parts of this section were experienced during the exceptionally wet winter of 2000/01, but quantitative estimates of additional sediment yield are not yet available.
In recent years there has been landslide movement at the sea road access to Barton 2009-2011 (West, 2013). The Channel Coastal Observatory is currently engaged in detailed monitoring of topographic changes, including cliff top edge recession and gulley enlargement, using ground survey and lidar. Topographic change is plotted as differences between successive surveys, with retreat of the cliff top evident from overlays on orthorectified aerial photography for four sorties between 2000 and 2008 (Picksley and Philip, 2011; Picksley, 2012). It is evident that the most discernible change in recent years has been in front of the Cliff House Hotel and at Hoskins Gap. Groundwater piezometer readings and rainfall data are also being continuously recorded.
The cliffs along this segment are composed of the predominantly sandy Upper Barton Beds and exhibit a steeper upper profile than further west. Nevertheless, lithological variations occur and the lower part of the cliff has high clay content resulting in a characteristic scarp and bench profile regulated by frequent mudflows. Historically, retreat averaged 0.85m per year over the period 1869-1959 (Hooke and Riley, 1987), although a more rapid retreat of 1.5m per year was recorded for the period 1932-68 (Lacey, 1985). Since approximately 1960, retreat has accelerated significantly to a mean of 2.0m per year, with between 3.6m per year and 4.2m per year recorded for the period 1968-82 (Nicholls, 1985; Lacey, 1985; Dixon, et al. 1998; Halcrow, 1999; Brown and Barton, 2007). This acceleration is primarily attributed to accentuated terminal scour (outflanking) resulting from reinforcement in 1970 of the Becton Bunny outfall and construction of a cliff toe revetment and sequence of rock groynes along the updrift Barton-on-Sea frontage (Photo 11) (Nicholls, 1985; Brown and Barton, 2007), and to related reduction of beach width protecting the cliff toe. The zone of cliff failure reactivation is migrating eastwards (Dixon et al., 1998; High-Point Rendel, 2002; Barton and Garvey, 2011). Cliff erosion supply was calculated at 15,000m³ per year (>0.08mm diameter) for 1984/85 which greatly exceeds that prevailing before the early 1970s (Lacey, 1985). For a 1.8km length of shoreline, Dixon, et al. (1998), calculated from high resolution aerial photography, 1967-1993, that total sediment yield for this period was 1,400,000m³ (approx. 52,000m³ per year). Immediately downdrift of Becton Bunny, the major morphodynamic zones of the cliff face shifted landwards, but an overall steady state balance of upper and lower profile processes was maintained. Nicholls (1985) calculated gravel input of 7,000m³ per year from cliff erosion west of Hordle, the majority derived from the Becton-Hordle segment, because the Becton outfall has operated as a partial, but effective barrier to input from further updrift.
Between Taddiford Gap and Hordle Cliff, retreat is slow, averaging 0.12m per year for the period 1809-1969 with negligible retreat recorded between 1931 and 1969 (Hooke and Riley, 1987). This is due to accretion of a substantial gravel beach (Photo 12), which protects the cliff toe and prevents marine erosion and fresh sediment input (Nicholls, 1985; Halcrow, 1999), and which is linked in an as yet undetermined way to a persistent nearshore bar below maximum low water (Halcrow, 1999). Royal Haskoning (2011) deduce the current (2004-2009) rate of recession at Hordle cliff to be 0.8m per year. Posford Duvivier (1997) calculate that the long-term yield of fine sediment has been approximately 57,000m³ per year from this sector, but that coarse sediment release is no greater than 1,500m³ per year.
The cliffs at Becton Bunny, either side of the terminal groyne are likely to provide a direct supply of eroded cliff material to the beach, continuing along towards Hordle. The clifftop along this section receded up to 10m 2005 to 2013 (aerial photography analysis). The E5 arrow volumes were originally >20,000m³ per year but this represents the potential total yield, which is not likely to be retained on the beach since the fine fraction is lost in suspension. The lower yield of 1,500m³ per year calculated by Posford Duvivier (1997) is used to support the 2012 updated E5 arrows.
Cliffs at gradients of between 20 to 45° composed of sands, clays and marls of the Headon Formation decrease in height from 20-25m at Hordle Cliff to beach level at Milford. Historically, Milford-on-Sea beach has shown a general trend of retreat. Map comparisons indicate a mean rate of retreat at 0.7m per year over the period 1869-1969, but increased recession was recorded during the latter part of this period, at a mean of 1.08m per year (Hooke and Riley, 1987; Nicholls, 1985). Since the early 1970s, erosion rates of up to 1.8m per year have been calculated (Halcrow, 1999) following up-drift cliff stabilisation and reduced beach volumes. Analysis of profile data at Milford-on-Sea between 1987 and 2006 calculated a retreat rate of 0.77m per year (Karunarathna, 2011). The cliffs are capped by Pleistocene gravels, attaining a maximum thickness of 7m at Rook Cliff. These are regarded as remnants of a formerly more substantial infilled tributary channel of the River Solent (Nicholls, 1987). Evidence is based on historical records of uncertain reliability and an observed trend elsewhere for inland thinning of gravel deposits. Parts of this cliff segment are colonised by vegetation, and have been protected by groynes and sea walls first installed in the early 1930s, upgraded in the late 1960s and protected by a fronting rock revetment in the late 1990s (Photo 12) (Dobbie and Partners, 1984; Halcrow, 1999; Royal Haskoning, 2011). Input supply of 3,422m³ per year (>0.08mm diameter) is calculated for 1932-68 compared with 3,000m³ per year for 1984/85 (Lacey, 1985). Posford Duvivier (1997; 1999) calculate a total yield of 12,000m³ per year (58% sand, 40% clay and 2% gravel). Continued supply, despite protection, can be attributed to rapid recession of unprotected parts, some continuing superficial instability of stable slopes and progressive loss of beach volume. The latter appears to have accelerated to a rate now exceeding 0.5m per year since the late 1980’s (Bray, et al., 1996; Bray and Hooke, 1998, Halcrow, 1999).
The E6 arrow volume was 3-10,000m³ per year in the 2004 version. However a total yield of 12,000m³ per year suggested by Posford Duvivier (1997; 1999) is most likely, 60% of which yields a potential 7,200m³ per year of beach grade material. The E6 arrow therefore remains at 3-10,000m³ per year.
Shoreface Erosion
Posford Duvivier and the British Geological Survey (1998) provide some very approximate calculations of wave and tidal current-induced erosion of the shoreface of Christchurch Bay. This feature has a width that is generally less than 1,000m, narrowing to less than 350m at Milford and is mostly developed in basement Eocene sands and clays. The following data is given:
The reliability of these figures is subject to considerable uncertainty, as they derive from the application of simplistic formulae to base data of undetermined accuracy.
Interpretation and Summary
Assuming that contemporary cliff geology is similar to that eroded previously, supply patterns can be reconstructed for the past 100-150 years. Overall cliff input in Christchurch Bay has declined from 63,000m³ per year (particle size >0.08mm diameter) for 1867-1932 to 44,000m³ per year for 1932-68. Although this must partly reflect stabilisation measures involving drainage, groynes and sea walls at Highcliffe, Barton and Milford, full cliff stabilisation was not achieved until the late 1960s and early 1970s. It is possible that removal of obstructions to littoral drift beneath Hengistbury Head between 1848 and 1870 resulted in increased input from Poole Bay, until feed was reduced by construction of the Long Groyne in 1938. This input is corroborated by large scale spit growth across Christchurch Harbour entrance in the nineteenth and early twentieth centuries. Thus it can be postulated that sediment could have been supplied to the central and eastern parts of Christchurch Bay by the early 1900s and caused diminution of cliff erosion between approximately 1910 and 1940. Modern cliff input is in the order of 20-40,000m³ per year (>0.08mm diameter) with the higher figure only slightly less than previous supply despite cliff stabilisation at Highcliffe, Barton and Milford-on-Sea. Unprotected cliffs are of soft material vulnerable to erosion and therefore are still a source of beach material.
- A variety of historical, onshore and offshore sedimentological evidence has been compiled in support of the hypothesis that gravel-filled Pleistocene channels approaching 18m in thickness previously outcropped along more seaward positions of the cliffline but have now been largely redistributed by coastal erosion. If true, this suggests that gravel supply may have been twice as great and sand supply 8-10 times greater than at present. Verification of this is difficult for the conclusive evidence (the deposits themselves) is now dispersed. Input of all lithological types was 136,000m³ per year for 1867-1932, decreasing to 84,000m³ per year for 1933-68. Between 48 and 54% of all cliff input comprises sediments too fine to remain on the beach, and which are transported offshore in suspension. Christchurch Bay cliffs are thus a major source of suspended sediments whose ultimate fate is unknown, but which is almost certainly external to this cell.
- Application of crenulate bay concepts to Christchurch Bay reveal that up to 240 m of further recession is required for a stable plan shape to develop. (Halcrow, 1999; Royal Haskoning, 2011). This analysis is based on idealised conditions, with an assumed initially straight shoreline, uniform beach materials, homogeneous geology and a constant dominant wave direction. Few of these assumptions are applicable to Christchurch Bay, so the validity of this argument remains uncertain, though conceptually valid.
2.4 Beach Nourishment
» N1 · N2 · N3 · N4
Beach renourishment and recycling operations have provided a significant source of sediment, and has been practised since the late 1970’s at several locations including Mudeford Spit, Highcliffe and Barton (May, 1990). Over the last decade, this practice is ongoing at Mudeford Sandbank, Highcliffe, Milford-on-Sea and Hurst Spit.
N1 Hengistbury Head to Mudeford Spit
Small-scale beach recycling has been undertaken since the late 1970s both to the immediate north-east of the Hengistbury Long Groyne (Hydraulics Research, 1986a) and along Mudeford beach (Tyhurst, 1985a), the latter involving three separate operations each of 1,000 tonnes of coarse sand. The Mudeford Sandbank Coastal Defence Strategy Stage 1 Works were completed during May 2000. Works included extension and heightening of existing rock groynes and provision of 'basement' nourishment under the sand foreshore (Wallingford, 2001).
In recent years, Christchurch Borough Council have conducted recycling operations through extraction of excess gravel and sand built up over time from the Mudeford Sandbank and other natural areas of accumulation to updrift groyne bays along Mudeford Spit. This is conducted as part of the Beach Management Plan for Mudeford Sandbank. Recycling activities (2003-2013) have been compiled from recycling logs and are summarised in Table 1 below. Locations refer to groyne numbers, with ‘S’ groynes along Mudeford Spit.
N2 Avon Beach to Highcliffe
Falling beach levels in the 1970s prompted an initial sand recharge, followed later by a sand and gravel nourishment scheme to compensate for ongoing volume losses. The artificial beach profile was designed by two-dimensional modelling in a wave flume which simulated storm wave and surge conditions (Hydraulics Research 1984, 1986b; Tyhurst 1985b). Optimum beach configuration was subsequently achieved using 55,000m3 chert and flint gravel from an inland source. Beach retention was facilitated by existing timber groynes and two newly constructed downdrift rock groynes. This was completed in March 1985 and was subject to the regular monitoring of 30 transverse profiles measured between October 1984 and June 1986. Analysis revealed a 14% increase in volume (6,700m³) for the period March 1985-February 1986, although the initial volume of the artificial beach was uncertain because its packing density was not measured (Hydraulics Research, 1986b). The measured volume increase may thus be partially an artefact of the monitoring technique. Further profile monitoring in February and March 1987 indicated an overall volume loss of 5% (Hydraulics Research, 1988). This information is difficult to compare with previous monitoring because volumes were calculated above Ordnance Datum, whilst the previous survey calculated volumes above -0.5m OD. If both surveys are accepted as accurate, net sediment gain of 6,700m³ is estimated for March 1985 to February 1986, followed by loss of 8,600m³ between February 1986 and March 1987. Profile monitoring and analysis up to June 1990 revealed diminution of beach volume from 57,130m³ (March, 1985) to 45,100m³ (above -0.5m ODN), a total loss of 21% (2,400m per year) over the five year period (Hydraulics Research, 1991). Accepting the measured gain in the first year followed by rapid loss in the second year as an expected result of renourishment, loss between March 1987 and June 1990 is estimated at 3,100m³ per year. Because the effects of compaction and the quantity of natural beach feed from onshore transport is uncertain, it is difficult to determine losses accurately, although the pattern of rapid initial drawdown followed by a decreasing rate of loss in subsequent years is characteristic of gravel nourishment schemes As a result of these losses, an improved scheme entailing replenishment with 18,000m³ of gravel and replacement of existing timber groynes with rock groynes was undertaken in 1991, with subsequent maintenance. Details of its performance are not available, but progressive loss of volume is reported (Coates and Bona, 1997). However, the rate of loss has been lower than that which is characteristic of most other south coast gravel recharge schemes.
More recently, Christchurch Borough Council have conducted recycling operations along the frontage between Avon Beach (Mudeford) and Highcliffe. Sand and gravel that naturally accumulate in some groyne bays are recycled regularly to sediment-starved bays. Recycling activities (2003-2013) have been compiled from recycling logs and are summarised in Table 2 below. Locations refer to groyne numbers, with ‘M’ groynes along Mudeford Quay to Avon Beach, ‘F’ groynes along the Friar’s Cliff frontage and ‘H’ groynes adjacent to Highcliffe.
Barton-on-Sea
Small-scale gravel renourishment has been carried out in connection with the construction of rock groynes in the mid-1980s, to provide bay-shaped beaches (Fleming and Summers, 1986; Wright, 1992). No further nourishment activities occurred 2003-2012.
N3 Milford-on-Sea
There is little information regarding the frequency or volumes of recharge at Milford-on-Sea prior to 2008, however it may have occurred since the mid-1970’s. Small scale recharges were conducted in the early summer of 2008 due to lowering of beach levels. In the late summer of 2008, emergency works were conducted in response to sea wall failure and promenade collapse at the western end. The works included rockwork and beach recharge of 15,000m³, completed in March 2010.
In more recent years, New Forest District Council have conducted small-scale recycling and renourishment operations along the Milford-on-Sea frontage in response to sustained lowering of beach levels. Sources of sediment include material extracted locally sourced quarry material. Two annual recharges, each of 5,000m³, were undertaken in 2011 and 2012. Recycling and replenishment activities (2003-2013) have been compiled from recycling logs and are summarised in Table 3 below. Locations refer to ‘beach profile’ numbers.
N4 Hurst Spit and North Point
Hurst Spit has been declining in volume, and its foreshore receding, since at least the late nineteenth century (Hooke and Riley, 1987; Bradbury, 1998). Rates of erosion have undoubtedly been accelerated by the interruption to, and reduction of, the rate of littoral transport due to updrift protection works in Christchurch Bay, whilst losses to the offshore zone at the distal end have been more or less constant over this period. The consequence has been increasingly severe erosion of this barrier structure, with overtopping, overwashing, crest cut-back and breaching occurring under conditions of storm surge (Bradbury, 1998; Bradbury and Kidd, 1998 - further morphological and morphodynamic description and explanation is given in Section 5).
To prevent breaching at the proximal end, a 600m length of rock armour was emplaced between 1967 and 1968. Renourishment at an annual average of 1,000m³ was undertaken between 1980 and 1985. Storm erosion in 1984 resulted in widening, reprofiling and recharge over a 450m length of the spit beyond the earlier rock armour (Dobbie and Partners, 1984). Severe storms in October and December 1989, with a 1:100 year recurrence, caused overwashing, crest flattening over some 800 m of the spit, up to 80m of "rollback" (landward migration), and the displacement of some 50,000 tonnes onto the leeward saltmarsh of Mount Lake. Unquantified offshore losses also occurred. The immediate response was to recover some 25,000 tonnes of gravel from Mount Lake, and import an equal quantity from inland sources, to rebuild the structure, now some 12 m set back from its pre-storm position. Some recharge was also carried out following major storm-induced erosion in February 1991 and April 1994 (see Section 4.3).
In response to the impact of the 1989 storms, and because imported recharge sediments from inland sources proved to be too small to be retained over the longer-term, New Forest District Council undertook in-house and commissioned a series of field-based and model investigations, together with ongoing routine monitoring of beach behaviour, to determine a long-term, optimum scheme of protection (HR Wallingford, 1993; Mackintosh and Rainbow, 1995; New Forest District Council, 1990, 1996; Wimpey Environmental Ltd, 1994; Bradbury and Kidd, 1998; New Forest District Council, 2010). The outcome, constructed in 1996, involved a series of measures (see Section 5.3 for further detail), which included the recharge of the most vulnerable 800m length of the proximal and mid-sectors of the spit (Photo 14). This was obtained from the nearby Shingles Bank offshore, as it is the natural source of supply of gravel to the spit and functions as a store for material removed from it. It thus provided a source whose size grading was very close to the prototype beach material. The effect of this recharge was to almost double the previous volume of the spit. Crest level and width were both increased, though declining eastwards in conformity to the reduction in wave climate severity from proximal to distal ends. Allowance was also made for subsidence, as the basement support of the spit is composed of low yield strength salt marsh deposits, over which it has migrated.
Most recent information from the Regional Coastal Monitoring Programme (Channel Coastal Observatory, 2008) indicates a net erosional trend on the lower foreshore along Hurst spit. Accretion is observed along the rear of the beach slope – this correlates with the sediment recycling exercises which have occurred during 2003, 2004, 2005 and 2008 and have mainly placed material along the rear of the slope (Royal Haskoning, 2011). Information on more recent sediment recycling and renourishment have been compiled using sediment recycling logs in Table 4 below. Natural accumulations of sediment at the end of North Point are recycled approximately every three years to areas along Hurst Spit which have been damaged during storms. Locations refer to ‘beach profile’ locations.
3. Littoral Transport
Introduction
Analysis of Regional Coastal Monitoring Programme lidar, aerial photography and topographic baseline survey data 2003 to 2012, combined with other datasets, academic research and historical studies has enabled sediment budgets, transport rates and directions to be identified or verified where possible.
Analysis of Coastal Monitoring Programme lidar, aerial photography and baseline topographic data for 2003 and spring 2013 confirms net eastward littoral drift within the bay. Beaches at the western end of the bay are predominately sandy, with proportions of gravel increasing from Barton-on-Sea further east, where Hurst Spit is composed primarily of shingle. The presence of nearshore bar features also adds complexity when determining littoral transport rate. The beaches at the eastern end of the bay are undergoing a net loss of sediment although a sediment budget is difficult to calculate because it is not possible to determine whether the sediment leaves the subcell or transported subtidally.
The overall net pathway of bedload transport is eastwards, with alongshore wave energy approximately proportional to drift rates. Wave approach direction is considered more important than inshore wave heights in determining spatial variations in longshore transport efficiency. Since the mid-twentieth century, the rate of longshore movement of sediments has been slowed by the presence of groyne compartments at Highcliffe and Barton. Cliff stabilisation and protection has significantly reduced the volume of sediment available for longshore transfer. The proportion of sand declines from west to east, indicating that finer materials are winnowed offshore. Drift supply to Hurst Spit is almost entirely of gravel.
(See also the relevant section in the account of Poole Bay)
Littoral drift at Hengistbury Head has varied over the past 160 years, mostly due to human activity. Between 1848 and 1870, ironstone boulders were mined from the foreshore at Hengistbury and from Christchurch Ledge. This is believed to have caused marked acceleration of west to east littoral drift, which was previously partially blocked by the larger salient form of the headland (Tyhurst, 1985a; Halcrow, 1999). Drift - especially of gravel - was again very substantially intercepted by construction of the Long Groyne in 1938 (Lelliott, 1989).
A substantial gravel and sand beach has accumulated to the west of the Long Groyne, indicating net eastward drift. Analysis of volumetric change at this barrier indicates drift of 600m³ per year (1968-78) to 900m³ per year (1938-68) but this only relates to material coarse enough to be retained (Webber, 1980b). Fluorescent sand tracer experiments on Solent Beach indicated net eastward drift of 30,000-50,000m³ per year per year but these figures are basically conjectural as they are unlikely to be temporally representative (Webber, 1980b). A wave refraction analysis based on Portland wind data for 1977 was employed to determine longshore wave energy flux. Use of appropriate sediment transport equations calibrated for sediment size enabled a littoral drift calculation (Henderson, 1979). A rate of 86,000m³ per year was calculated at the Long Groyne, with a potential rate accelerating to 645,000m³ per year northward to Mudeford Spit (Henderson, 1979). Similar analysis using improved sediment transport equations indicated drift of 96,000m³ per year to the immediate west of the Long Groyne and 575,000m³ per year northward to Mudeford Spit (Lacey, 1985). A further improved longshore wave energy flux technique suggested drift of 45,000m³ per year towards the Long Groyne (Hydraulics Research, 1986a). Several conclusions can be drawn from these estimates. The wide range of figures not only suggested that some of the analytical techniques were subject to substantial error but that different approaches measured different attributes. Beach volume analysis primarily measured coarse sand and gravel accumulation, whilst wave energy flux calculations were biased towards sand moving as both bedload and suspended load. However, it is apparent that the littoral drift rate accelerated markedly north-east of the Long Groyne, although it is probable that the values of Henderson (1979) and Lacey (1985) are large overestimates. This is because the refraction technique performed poorly in areas of complex bathymetry, e.g. Christchurch Ledge, and did not adequately model the local diffracting/reflecting effects of the Long Groyne (Hydraulics Research, 1986a) and therefore is of low reliability.
The barrier effect of the Long Groyne is therefore difficult to assess. Some authors suggest that it prevents littoral drift (e.g. Wright, 1982), whilst others state that it is only a partial boundary (e.g. Tyhurst, 1985a). Limited evidence available favours the second explanation, with almost total interruption of gravel transport, but only partial interception of sand. A physical model study of the Long Groyne and co-adjacent areas at 1:80 scale employed 1.5mm diameter coal particles and 0.37mm sand grains to simulate natural (prototype) gravel and sand respectively. Operation of the model over a variety of representative wave conditions accurately replicated observed beach behaviour; furthermore, perspex in suspension was carried eastward by longshore currents and was pushed offshore by the Long Groyne. Transport was eastward or east-south-east towards Christchurch Ledge, with no evidence of a return (onshore) feed to Mudeford Spit. By contrast, no coal particles were carried around the Long Groyne (Hydraulics Research, 1986a). Quantitative information could not be obtained as it was not possible to test a sufficient number of wave conditions, but mathematical model studies indicated net eastward or north-eastwards longshore bedload movement of approximately 45,000m³ per year (Hydraulics Research, 1986a); actual sand bypassing the Long Groyne may be at a similar order of magnitude. Reliability of this conclusion is uncertain because problems of model scale cause difficulty in selecting experimental material that is representative of the size and density of natural sediment. However, ability of the model to simulate observed beach changes probably indicates overall results to be of low reliability and therefore the reliability has been updated from medium to low reliability. The model predictions broadly agree with previous conjecture and some field observations, e.g. it was reported that eastwards sand flow around the Long Groyne fed Highcliffe rather than Mudeford (Wise, 1959). This was generally supported by the results of sea-bed drifter studies, which indicated a net offshore tendency west of Hengistbury Head, north-eastward transport over Christchurch Ledge towards Mudeford Spit and marked onshore transport between Mudeford Quay and Highcliffe (Tyhurst, 1976; Turner, 1990).
Rock groynes constructed on Mudeford Spit between 1988 and 1990 and upgraded in 2000/01, rapidly filled with sediment (Photo 15) (Tosswell, 1978; Tyhurst, 1987; Nicholls and Wright, 1991; Halcrow, 1999; HR Wallingford, 1999b). Historically, this location experienced a highly dynamic phase of spit extension and punctuated breaching between the mid-1800s and 1938, which has been variously attributed to onshore transport (Robinson, 1955) and littoral drift (Tosswell, 1978) operating together (see Section 5.2). By assuming that subsequent spit regression resulted mostly from abruptly reduced littoral drift, a rate of 107,000m³ per year has been calculated for the preceding period of most rapid documented change, i.e. between 1932 and 1939 (Lacey, 1985). Based on a limited experiment using aluminium pebbles, the bedload littoral transport rate immediately south of Christchurch Harbour entrance was computed to be 53,000m³ per year (Nicholls and Wright, 1991). H R Wallingford (1999b) calculated the prevailing rate to be in the order of 40,000m³ per year, based in part on numerical modelling of wave climate modified for local wave refraction. Most of this takes place in the nearshore zone with sand interception by the present system of rock groynes providing an estimated beach drift rate not greater than 16,000m³ per year. (Halcrow, 1999).
It was therefore concluded that a significant difference in littoral drift potential existed either side of the Long Groyne. West of the groyne, drift has been estimated by a variety of techniques and an approximate order of magnitude established, with low reliability. To the east, drift estimates are affected by complex refraction and diffraction effects. Despite this, a sharply accelerated rate of drift is indicated. Aeolian transport of sand along and across Mudeford Spit has been largely neglected, although HR Wallingford (1999b) estimate a gross rate of approximately 11,000m³ per year, with losses of some 9,000m³ per year due to onshore and offshore movements; the net longshore throughput is therefore no more than 2,000m³ per year.
The Long Groyne is an effective impediment to littoral drift of gravel, although limited quantities may overtop the backshore part of this barrier during storms (Hydraulics Research, 1986a). Sediment shortage is the main contributory factor to low rates of gravel movements with the sources of supply for this pathway uncertain. Onshore transport from Christchurch Ledge had been proposed (Tosswell, 1978) and may be supported by diver observations of mobile sand and gravel (Collins and Mallinson, 1986; Gao and Collins, 1994c, 1995a). Halcrow (1999), however, infer that gravel is likely to be retained by the 'troughs' developed in the seabed bedrock outcrop of Christchurch Ledge.
Analysis of Coastal Monitoring Programme lidar, aerial photography and baseline topographic data was used to support elements of the historic studies into sediment transport around the Long Groyne and along Mudeford Spit for the period 2003-2013. Several factors contributed to difficulty in calculation of a sediment budget; interaction between the beaches and ebb tidal delta which is difficult to quantify, the low reliability of input volumes around the Long Groyne, fluctuations in sediment transport rate due to updrift major renourishment schemes, and sediment recycling between the ebb delta and beaches along Mudeford Spit.
Use of baseline topographic difference models 2003-2012 suggests low beach volumetric net change per year, with no evidence of accumulation to support high volume sediment transport rates. There is a notable increase in volume 2006 to 2007 along Mudeford Spit, attributed to a large Bournemouth replenishment scheme updrift during winter 2005 and 2006. This highlights the variability in sediment transport supply and consequent fluctuations in sediment transport rate over time.
The earlier research does not yield quantitive sediment transport rates of high reliability that can be compared with recent measurements. Many studies were conducted prior to construction of rock groynes along Mudeford spit, and do not account for recent beach recharge and recycling schemes in Bournemouth and Christchurch Bays. The LT1 arrow therefore has been amended from >20,000m³ per year to “no quantitative data”.
This shoreline sector is suggested as a littoral drift sub-cell, because prior to major protection in the late 1960s, marked accretion was recorded in the vicinity of Highcliffe Castle. By contrast, the segment immediately west of Chewton Bunny exhibited erosion, so the original sub-cell boundary was located between these areas (Nicholls, 1985). A similar distinction is made by Lacey (1985) based on analysis of wave energy flux, and attributed to variable offshore bathymetry, which causes complex wave refraction and longshore variation in littoral drift rate potential (Henderson, 1979; Lacey, 1985). It must also be recognised that coast protection structures intercept drift, generating littoral drift boundaries. Construction of groynes at Highcliffe in and after 1971 intercepted littoral drift and therefore created new transport compartments at either end of the protected section; this was further enhanced by beach nourishment and construction of rock groynes further east in 1985. Subsequent monitoring of the beach involving measurements at 30 profiles over a 5 year period revealed a significant eastward shift of beach sediment identifiable in the first year (Hydraulics Research 1986b) and continuing in subsequent years (Hydraulics Research, 1988; 1991). It was concluded that eastward drift was mainly responsible for the measured diminution of the nourished beach. The groynes were therefore not a total barrier to littoral drift but only intercepted sediment on the upper beach. Extended beach profiles showed that the nearshore and offshore zones were dynamic and it can therefore be postulated that offshore-onshore transport of sand during storm-swell conditions also involves a net eastward transport component which carries sediment around these structures (Lacey, 1985). A mean drift rate could not be determined from beach profile measurements because inputs to the protected segment were not measured and nearshore-offshore information was limited. Calculations based on longshore wave energy flux and transport equations calibrated by measurement of sand loss on a replenished beach (Bournemouth) indicated an eastward potential drift of 191,000m³ per year (Henderson, 1979) and 195,000m³ per year (Lacey, 1985). These estimates may be unreliable due to difficulties with the refraction technique over the complex bathymetry of Christchurch Bay (Henderson, 1979). Despite these problems, drift direction is correctly predicted according to observations by Robinson (1955), Tosswell (1978) and Tyhurst (1987), and all analysis indicates that drift is less rapid than between Hengistbury Head and the entrance to Christchurch Harbour.
The variable sediment sources and the difficultly in assessing the volume of sand supplied to Mudeford Sandbank combine to make longshore transport rates difficult to predict in this area. However, past experience combined with modelling provides estimates total along the frontage of 50,000m³ per year longshore transport (Wallingford, 2001). It should be noted though that this may be an overestimate since wave attack along the frontage is particularly oblique and supply of sediment at Hengistbury Head is likely to be limited. Much of transport that does take place occurs below the low water level. This area is beyond the extent of the Coastal Monitoring Programme data.
Littoral drift input from Mudeford Spit to the Mudeford - Highcliffe sub-cell is possibly achieved by two mechanisms. Between the mid-1800s and early 1900s, growth of Mudeford Spit caused periodic lengthening of "The Run", building up a hydraulic gradient across the spit and facilitating breaching (Burton, 1931; HR Wallingford, 1999b; Royal Haskoning, 2011). The detached spit was then driven onshore to supply Mudeford and Highcliffe (Robinson, 1955). This process has not been operative since the 1930s, although it is believed that bypassing of the inlet is facilitated by sand drift via storage on an offshore bar (Tosswell, 1978 - see Section 5.2 for further discussion). The relatively healthy sandy beaches at Mudeford (Photo 16) and Friars Cliff (Photo 3) are indicative of a process of continued onshore supply otherwise they would become depleted by the west to east net drift.
The frontage between Avon Beach (Mudeford) and Chewton Bunny is subject to sediment recycling from areas of natural accretion, to areas of noticeable erosion as detailed in Section 2.4. This complicated the production of a sediment budget. It should also be noted that a proportion of sediment transport is likely to be occurring below MLWS in the nearshore zone where there are few quantitative measurements.
The previous SCOPAC (2004) STS had calculated the sediment transport rates to be >20,000m³ per year along this section based on the literature review. In summary, the text concludes that quoted rates are likely to be an over-estimation and are ‘potential’ sediment transport rates. The impact of coastal defences on sediment transport rate implies that sediment transport rates are much lower than ‘potential’ sediment transport rates.
Analysis of Coastal Monitoring Programme lidar for 2008 and spring 2013 difference plots shows notable erosion in the western end of this frontage, switching to accretion beyond the groynes at Friar’s Cliff towards the groynes at Highcliffe. Some groyne bays, and areas above MHW along the Highcliffe to Chewton Bunny section, demonstrate natural accumulation, which is a source of sediment used by Christchurch Borough Council for recycling. The 2013 aerial photography shows potential for interaction between the beach and nearshore zone, with sand bars and accumulations of sand visible in this area.
Further work that studies the interaction of the beach and lower foreshore below MLWS are recommended, but is beyond the current scope of the Coastal Monitoring Programme. As a result of the discussion above, the volume for LT2 has been amended from >20,000m³ per year to “no quantitative data”.
Construction of groynes and a timber revetment at Barton-on-Sea as part of the 1964-69 protection schemes divided the shoreline into two littoral drift sub-cells: a natural western sub-cell at Naish with unimpeded transport and an eastern sub-cell at Barton-on-Sea subject to interrupted transport by groynes (Nicholls, 1985; Bray and Hooke, 1998). Compartmentalisation of the beach has subsequently been increased by construction of four rock groynes/strongpoints completed by 1990.
West Barton beach was monitored by intensive cross-section profiling involving 47 surveys along 12 survey lines over the period 1976-78 (Webber, 1980b; Halcrow, 1980). These revealed net accretion of 16,000m³ per year over the period, but probably were not representative of long-term trends because net intertidal erosion of 3,000m³ per year and nearshore erosion of 20,000m³ per year was recorded from a larger data set covering the period 1974-82 (Lacey, 1985). These data are not easily converted to littoral drift volumes because these processes may operate with no change in beach volume. Estimation of drift is possible by integration of all known inputs and outputs, but has not been undertaken for this sector. Profile analysis of the protected East Barton beach revealed intertidal accretion of 2,000-4,800m³ per year and nearshore accretion of 3,000-63,000m³ per year over the period 1976-84 (Lacey, 1985). Although it is difficult to determine longshore transport, a distinct tendency exists for net beach erosion in the western (unprotected) part, a sector affected by the terminal groyne at Chewton, and accretion in the east where groynes intercept material (Mackintosh and Rainbow, 1996). Profiles indicated that seasonal onshore-offshore exchanges of sand and gravel (with net onshore transport during the summer and net offshore transport during the winter) were much greater than longshore changes (Halcrow, 1980; Webber, 1980b). Additionally, it was apparent that the sand component was more dynamic.
Sand transport was examined by fluorescent tracer experiments in 1976 and 1977 (Babbedge, 1987a and b; Webber, 1980b; Halcrow, 1980). Reliable drift estimates were only obtained for the first two tides of the main experiment, thereafter the technique proved inaccurate. These rates were combined with mathematical calculation of longshore wave energy flux to calibrate sand transport equations, and were subsequently used to calculate net drift from an annual wave climate. Drift volumes of 109,000m³ er year and 220,000m³ per year were obtained with the first and second tides respectively. These are not representative annual drift volumes because experimental data covered very restricted wave conditions untypical of overall wave climate. Better estimates of littoral drift were obtained by including refraction effects to generate longshore wave energy flux and combining these with previously established sediment transport calibrations. Analysis of this type for west Barton indicated drift of all sediment types combined between 167,000m³ per year (Henderson, 1979) and 308,000m³ per year (Lacey, 1985), the latter employing an improved sand transport calibration. Rates were determined at regular intervals along the coast (i.e. not based on natural transport boundaries), thus the drift rates quoted are applicable from just west of Chewton Bunny to the approximate boundary between Barton West and East beaches. Much lower drift rates of between 2,000m³ per year (Henderson, 1979) and 21,000m³ per year (Lacey, 1985) were determined for the sector covering East Barton beach to Becton Bunny. This was attributed to diminished longshore wave energy flux due to a local refraction effect, and lower drift efficiency due to increased sediment retained on the beach. Although incomplete understanding of the relative importance of factors and processes precludes reliable evaluation, the predicted eastward reduction of drift rates can be corroborated by visual evidence of beach and nearshore accretion. Cliff input analysis suggested that gravel drift past Becton Bunny may have been 7,600m³ per year for the period 1867-1971, thereafter diminishing to almost zero with outfall modification (Nicholls, 1985). This figure is within the range of the energy flux derivations and provides a fair degree of corroboration.
Towards Becton Bunny at the eastern limit of this unit, eastward drift of sand and gravel was substantially intercepted by an outfall between 1971 (when the outfall was reinforced and a rock and sheet-pile strongpoint was constructed) and 2001. In 2003, the outfall was removed from this location and diverted so that it now discharges 600m to the west. The rock structure remained, and therefore continues to partially intercept littoral drift. At the new outfall location, the associated rock structure also appears to intercept eastward littoral drift.
This section is now considered as two different units. There is the western undefended ‘Naish’ beach with naturally-eroding cliffs to provide sediment input. The eastern section at Barton-on-Sea is currently protected with a series of rock revetments and groynes, which impede sediment transport. Neither section has been subject to recycling or replenishment over the last decade.
Analysis of Coastal Monitoring Programme lidar data for 2008 and spring 2013 demonstrates accretion along the Naish section. There is a nearshore sand bar extending from the western end that interacts with the beach, making a sediment budget difficult to calculate. The Barton-on-Sea frontage no longer receives natural cliff input due to the coastal protection works that effectively separate the beach from the cliffs, apart from in the aforementioned locations. The beaches between the rock groynes are narrow above MLWS, with sediment moving from west to east within the bays. There is little change to beach volume 2008 to spring 2013 towards Becton Bunny. Further work is required to understand the contemporary sediment transport rate along this frontage that combines the beach and lower foreshore below MLWS.
The previous SCOPAC (2004) sediment transport study estimated the sediment transport rate for LT3 to be 10-20,000m³ per year. This was based on the literature review. The literature is not able to define a recent sediment transport rate since the last major coastal protection works were completed in 1990. It is likely therefore that recent sediment transport rates are much lower due to the impedance of sediment transport rate caused by the rock groynes and revetment. The LT3 arrows are therefore “no quantifiable data”.
Estimates of littoral drift using the energy flux technique range from 2,000m³ per year (Henderson, 1979) to 40,000m³ per year (Lacey, 1985), the difference resulting from sediment size calibration; the lower rate assumes an all-gravel beach, as opposed to a 54:46 sand: gravel beach used for the upper rate. Theoretically, the latter should be more accurate because sediment size calibration was based on beach sampling (Lacey, 1985). An alternative method based on historical changes of the beach, cliff and nearshore zone determined the minimum littoral drift necessary to produce the measured change. This analysis showed that gravel drift diminished from between 8,400-13,800m³ per year for 1939-68 to 7,000m³ per year for 1969 to 1982 (Nicholls, 1985). The reduction was attributed to the Barton Cliffs protection programme and interception of drift by the Becton Bunny outfall. Long-term beach gravel accretion in the form of a small cuspate foreland with a volume between 0.5 and 1.5 million m³ is recorded at Hordle Cliff from map comparisons, at rates of 2,600-2,900m³ per year for 1867-1939, 8,400m³ per year (1939-68) and 5,900m³ per year for 1958-82 (Nicholls, 1985; Halcrow, 1999). This location is widely recognised as a transport discontinuity, with accretion occurring as the littoral drift rate reduces to near zero (Nicholls, 1985; Nicholls and Webber, 1988a; Halcrow, 1999). The effect is believed to be due to local wave refraction which causes a relatively low longshore wave energy flux (Nicholls, 1985). Map analysis indicated that this accretion zone migrated 500m to the west between the periods 1867-98 and 1938-68, reflecting supply from the west at a more rapid rate than output to the east. After 1968, this differential was reduced; westward migration ceased and a reverse trend for eastward migration was initiated (Nicholls, 1985). Beach profiling between 1981 and 1982 revealed a loss of 24,000m³ of predominantly coarse sandy sediments at Hordle, but it was concluded that survey duration was too limited to distinguish any long-term trend (Nicholls, 1985). Halcrow (1999) indicate that beach erosion has been prevalent, albeit at a slow rate, since the early 1980s.
This section receives a natural source of sediment from the undefended cliffs between Becton Bunny and Hordle. This input is mostly fine sediment, which is not retained on the upper beach above MLWS. It may however contribute towards the sandbar which extends from the terminal groyne at Becton Bunny 3km east towards Milford in the nearshore zone, clearly visible in the 2005 and 2008 aerial photography. The position of this bar varies over time, with some sections interacting with the beach. This bar is not captured consistently within the topographic data, as it varies in position below MLWS. The beach above MLWS is narrow at the Becton Bunny end, widening towards the east with increasing proportions of gravel.
The western limit of this section is considered to be a partial sediment transport boundary due to the installation of rock groynes along the Barton-on-Sea frontage. The beach at Hordle exhibits a net loss of material from the upper beach identified using a lidar difference model 2005 to 2012. This may be loss of fine sediment from the cliffs which falls onto the beach and is removed to the bar, as the erosive beach trend peters out further east as the cliffs stabilise.
The previous SCOPAC (2004) STS estimated the sediment transport rate for LT4 to be 10-20,000m³ per year. This was based on the literature review. The literature review is not able to quantify a more recent sediment transport rate and is therefore “no quantifiable data”.
The Hordle to Milford section is backed by naturally eroding cliffs, although the supply to the beach is mostly restricted by beach huts. Along the Milford section, the beach is interrupted by groynes and rock revetments. The beach is composed mostly of gravel along this frontage.
Historically, the Hordle drift boundary has not acted as an absolute barrier to littoral transport, but its effect has been progressive, allowing 65% of longshore movement to pass in 1867-98, 61-62% in 1908-1939 and 0-36% in 1939-68 (Nicholls, 1985). Due to the intercepting effect of this foreland, drift between Hordle and Milford diminished from a maximum of 4,900m³ per year during 1939-68 to 1,400m³ per year between 1969-82 (Nicholls, 1985). It is reported that timber groynes have been present on the Milford frontage since 1937; three rock groynes or strongpoints were constructed at the eastern extremity in 1983 (Dobbie and Partners, 1984). It is likely that these structures have been effective in intercepting material so that gravel transport downdrift of Milford (i.e. towards Hurst Spit) is now extremely limited (Nicholls, 1985) and is therefore “no quantitative data”. Longshore wave energy flux analysis indicated potential drift of 7,000m³ per year of gravel in the west, reducing to 2,000m³ per year at Milford (Henderson, 1979) and 71,000m³ per year of both sand and fine gravel reducing eastward to 21,000m³ per year (Lacey, 1985). The higher values are theoretically more accurate, but their discrepancy with historic rates suggests three possible sources of error: (i) Henderson's analysis was only for 1977, which may be unrepresentative of long term conditions; (ii) wave refraction analysis was inaccurate over complex bathymetry and (iii) the sediment transport calibrations employed are not suitable for this site, with Henderson (1979) failing to fully represent the movement of sand. This is clearly a larger volume than gravel.
Analysis of coastal monitoring data lidar for 2005 and 2012 shows a net loss from these beaches.
Littoral drift has been studied on Hurst Beach using a variety of techniques which have included historical planform changes, beach profile analysis, aluminium pebble tracer experiments and numerical modelling (see Section 5.3). A 500m segment at the landward end (proximal sector) of the spit opposite Sturt Pond has been protected by rock armour since the late 1960s. The effects of this protection reduced natural drift because of interference with longshore energy flux (Bradbury, 1998), and was further modified in 1996 when this armouring was reinforced and its gradient reduced as part of a comprehensive protection scheme.
Analysis of historic changes by map comparison yielded eastward drift of 7,500-18,300m³ per year over the period 1867-1968 (Nicholls, 1985). Beach profiles surveyed on 27 occasions over the period 1980-82 indicated drift of 11,000-13,000m³ per year along Hurst Beach, increasing to 13,000-15,000m³ per year at the distal end (Hurst Castle) (Nicholls, 1985). Direct measurement of littoral drift was undertaken by means of aluminium tracer experiments. Quantitative analysis was difficult due to low tracer recoveries (1-25%) caused to some extent by burial and rapid longshore transport, but especially by inferred offshore transport (Nicholls, 1985, Nicholls and Webber, 1987b). Reliable results were only obtained for a 7 day period before recoveries diminished to unacceptably low levels. Wave conditions over this period were insufficiently representative for accurate calibration of transport equations. Experiments were therefore of limited use for volumetric calculation of drift, but yielded valuable information on particle sorting. These were most clearly sorted according to their 'C' axes, with larger material preferentially transported either towards Hurst Castle (distal end) or moved down the beach face to the lower foreshore and nearshore areas (Nicholls, 1985; Nicholls and Webber, 1987b).
Net longshore transport for Hurst Spit has been estimated by modelling to be in the range of 10-15,000m³ per year. Field observations have been used to validate this and suggest that these are an over estimate (Bradbury et al., 2008; Bradbury et al., 2009). Losses of about 11,000m³ per year occurred for a period of 4 years following the major sediment replenishment in 1996. The average loss over 11 years was 7,300m³ per year. Survey data following storm events highlight episodic periods of sediment drawdown (>40,000m³) temporarily deposited below mean water level (Bradbury et al., 2009).
Longshore wave energy flux determined for Hurst Beach by wave refraction analysis predicted westward net drift (Henderson, 1979). This contradicts reliable field observation and historical information and refutes the long accepted theory of spit elongation by eastward feed (Lewis, 1938; King and McCullough, 1971). This analysis is therefore clearly in error, probably resulting from failure of the refraction analysis to accurately predict waves travelling over the complex bathymetry of the Shingles Bank, North Shoal and North Channel. A more refined refraction analysis was employed to determine the inshore wave climate (Halcrow, 1982). Littoral drift was calculated from longshore wave energy flux, using adjustments to simulate variable tidal range and differing transport threshold conditions for the various sediment sizes. This technique yielded a net eastward drift of 15,000m³ per year. Brampton (1993), however, calculated that the mean drift rate for 1974-1990 was slightly over 36,114m³ per year (with a maximum of 51,600 and a minimum of 17,000m³ per year). This work was based on numerical modelling of the inshore wave climate, using a bulk sediment transport formula applied to monthly averaged wind speed data input into a wave hindcasting approach. Highest monthly rates, at 14,000m³ per year, occurred consistently in February of each year. No evidence for drift reversal was detected. These figures are considered an exaggeration of actual rates, as they are essentially based on theoretical reasoning and do not fully take into account the fact that wave climate varies along the length of the spit. Modelling undertaken for the 1996 recharge scheme concluded that a longshore transport rate of 16,000m³ per year was a reasonable estimate, but measurements following its completion revealed this to be too high. Between 1997 and 2003, rates downdrift from the offshore breakwater introduced as a component of the 1996 protection scheme averaged 9,000m³ per year, falling to 5,000m³ per year, 2004-2008 (Bradbury, et al., 2009). Updrift of this location drift rates are very low, if not close to zero. This reduced rate of littoral drift is arguably the result in part of subtle changes in the alignment of Hurst Spit following the modifications of the 1996 protection scheme, creating a more swash aligned orientation. More relevant is the fact that calculations of drift rates based on measured waves (available from the Milford Waverider buoy since 2003) are more reliable than earlier calculations based on proxy data. Refer to section 5.3.3 (Hurst Spit) for further discussion.
The wide range of techniques used to determine drift on Hurst Beach have produced results that established the sediment transport pathway with high reliability, but there is considerable uncertainty regarding actual rates of sediment transported. Therefore the LT6 arrow has been reduced in the 2012 update from >20,000m³ per year to “no quantitative data”.
Much of the predominantly coarse gravel moving northwards between Hurst Castle and North Point is lost offshore to Hurst Narrows, but a small quantity is transported northeast around the distal point and westwards along the recurve. The drift rate is estimated to be 2,000 to 2,700m³ per year at Hurst Point, declining to 900m³ per year at the northern recurve tip (Halcrow, 1982; Nicholls, 1985).
Accumulations of sediment at the distal end of North Point are extracted regularly and recycled locally in order to repair sections of Hurst Spit. This occurs on average every 3 years, and is of 5,000m³ per recycling event. This would suggest a transport rate of 1,500m³ per year. The 2004 volume was 3-10,000m³ per year, however this has been reduced to 1-3,000m³ per year.
Summary
- Net drift is clearly eastward along Christchurch Bay. A number of littoral drift sub-cells are present, but these do not represent self-contained sediment circulation systems because some transport takes place between adjoining sub-cells. Boundaries of these units are identified by significant changes in rates of littoral drift and by beach accretion zones at Highcliffe and Hordle. Development of principal sub-cells is attributed to the irregular bathymetry of Christchurch Bay which causes complex wave refraction and longshore variations in wave energy flux. Boundaries are variable and subject to longshore re-location, as at Hordle. Additional transport compartments have been created by protection of coastal segments by groynes, strongpoints and other structures which partially intercept littoral drift, e.g. Highcliffe, Barton and Milford.
- The western part of the bay is characterised by rapid transport of predominantly sandy sediments, towards the east. Sand is also mobile in cross-shore directions causing development of distinctive storm (offshore transport) and swell (onshore transport) profiles. Nearshore transfers frequently have a net longshore component, allowing quantities of sand to bypass coast protection structures. By contrast gravel is generally retained on the upper beach, so that littoral drift of coarse sediment is more effectively intercepted. Results of tracer experiments and beach profiling suggests that the presence of gravel on a mixed beach restricts the littoral drift of sand, as at Highcliffe, and may partly explain the relative success of beach nourishment there.
- The most comprehensive analysis of drift rates has been that based upon wave refraction studies and beach sampling, using empirical formulae to predict drift rates. Problems were encountered modelling wave refraction over the complex bathymetry of Christchurch Bay and the calibrations employed in sediment transport equations were of uncertain reliability. The earlier results are clearly in error for Hurst Spit and alternative estimates of drift rates based on long-term historical changes indicate both significant under- and over-estimation. Hence these results are only sufficiently reliable to indicate general trends and relative orders of magnitude.
4. Sediment Outputs
Three types of output can be recognised:
- From the beach system to various near and offshore sediment stores and/or sinks in Christchurch Bay.
- Permanent output from Christchurch Bay.
- Output from Christchurch Harbour.
These can be further distinguished according to the dominant transport mechanism. Estuarine output occurs via rapid tidal currents generated at the entrance to Christchurch Harbour and at Hurst Narrows. Offshore transport also occurs due primarily to wave action, but with some possible auxiliary contribution from tidal currents.
4.1 Transport in the Offshore Zones
O1 Shingles Bank
Research into sediment mobility has utilised: (i) side-scan sonar and echo sounding to map bedforms; (ii) sediment sampling, to infer pathways of movement; (iii) self-generated noise of gravel particle collisions, and (iv) measurements of near-bed tidal current and wave-induced stresses (Dyer, 1970; Velegrakis and Collins, 1992, 1993 and 1994; Velegrakis, 1994; Voulgaris, Workman and Collins, 1999). In addition, there has been some limited use of numerical modelling to derive information on probable sediment movements by waves and residual tidal currents (HR Wallingford, 1994; Brampton, et al., 1998; Halcrow, 1999).
All researchers identify strong offshore-directed (i.e. north-east to southwest) movement of sand and gravel in the Needles Channel due to the dominant ebb tidal current. The effectiveness of tidally-induced scour is demonstrated by the sharp boundary between the Chalk platform north of the Needles headland and the adjacent channel. In places, this sweeps clean the bedrock surface (Velegrakis, 1994). Gravel waves, with sharp crests, have an asymmetry that clearly indicates sustained south-westerly transport of coarse material along the easterly flank of the Shingles Bank (Velegrakis, 1994). The movement of sand in the same direction was deduced by Dyer (1970) from patterns of sand ribbons. HR Wallingford (1994) identified a separation of the south-west moving ebb current, with strongest flow (an average of 1.8m per second just above the seabed) along the western boundary of the Needles Channel. This was derived from application of TELEMAC, a depth-integrated finite element predictive model.
Sediment movements over the Shingles Bank are considered to result from stresses due to both waves and tidal currents, with waves effective to depths of at least 8m. Planar areas of the crest of the bank may be due to wave abrasion, probably under storm conditions; Velegrakis and Collins (1993), however, suggest that they are gravel-armoured surfaces from which sand is winnowed by tidal currents. This would be a year-round effect, thus the presence of sand patches overlying gravel - which appears to be confined to the winter months - indicates wave transport of sand. It is possible that tidal currents effect some abrasional scour, particularly on the eastern flank of the Shingles Bank (Velegrakis and Collins, 1993; 1994). Nicholls (1985) reported that parts of the crestal area build up to above maximum low water during periods of moderate wave energy, and are lowered during storms - i.e. sediment is dispersed under high wave energy conditions. Analysis of bathymetric and volume changes in the area of the 1996 dredging led to the conclusion that most transport was induced by waves (New Forest District Council, 1997-2001).
Voulgaris, Workman and Collins (1999) undertook an experiment that measured self-generated noise resulting from mutual collisions of gravel particles (median grain size of 1.70cm diameter) at the seabed, at a mean water depth of 8m. This was conducted over six consecutive days in an area near the north-western edge of the Shingles Bank, with well-developed gravel waves. Bedload transport rates remained nearly constant despite periodic fluctuations in the velocity of tidal currents, and variations in prevailing wave height. Waves were therefore considered to be the primary transportation mechanism, with transport rates correlating closely with particle weights. The relative dominance of wave action is likely to increase with water depth, i.e. to the south and south west.
Historically, erosion has been the dominant trend along the western flank of the Shingles Bank, with accretion (progradation) of its eastern margin (Velegrakis, 1994). Thus, a broadly west to east transport pathway has prevailed, involving the movement of both gravel and sand. Velegrakis and Collins (1993; 1994) report that the eastern flank exhibits an eastwards-thinning sediment wedge containing fine to medium sand some 10-15m in height. This is better-sorted than the sand fraction present along the western flank, and may indicate selective transport by tidal currents. Waves are considered the principal mechanism for moving coarse sand and gravel along this pathway.
Velegrakis (1994), revealed ambiguous evidence for net northwards or north-westwards movement of both sand and gravel along the western flank and in the area of North Head Shoal in the early 1990s; this was confirmed by New Forest District Council (1997-2001) on the basis of repeated bathymetric surveys which show the emergence of a distinct extension to North Head. Bedforms in this area lack any "steady state" asymmetry, tending to reverse with tidal current direction (Velegrakis, 1994). Thus, a combination of both tidal and wave-assisted transport is implied. North Channel was considered by HR Wallingford (1994) to be a product of tidal scour, though maximum (ebb stage) current velocities do not exceed 0.9m per second.
Two swath bathymetry surveys provide 100% coverage of Christchurch Bay. Through the Southeast Regional Coastal Monitoring Programme, the inshore 1 km zone of Christchurch Bay was completed in 2010. Further offshore, Poole and Christchurch Bays were surveyed in 2012 as part of the Maritime and Coastguard Agency’s UK Civil Hydrography Programme.
The swath bathymetry within Christchurch Bay highlights the complexity of the offshore sediment transport through the Shingles and Dolphin Banks. Starting at the Shingles Bank, which extends southwest from Hurst Spit, bedforms can be seen running along its entire eastern side. The bedforms indicate a lateral flow of sediment southwest along the bar. In the channel, bedforms can be identified in the same orientation, running from northeast to southwest, along the Needles Channel illustrating sediment leaving the West Solent (EO2).
O2 North Channel to Hordle: west-central Christchurch Bay
Asymmetrical gravel waves recorded by sediment sampling and side-scan sonar on the floor of North Channel indicate net north-westwards transport (Dyer, 1970; Velegrakis, 1994). A drifter experiment (Clarke and Small, 1967) also suggested movement in this direction, as well as onshore; however, other drifter studies (Watson, 1975; Turner, 1990) have given contrary results, leading to some confusion. Velegrakis and Collins (1994) note the sharp boundary between the western flank of the Shingles Bank and the adjacent area of the sea bed of Christchurch Bay. This may be the result of efficient "sweeping" by tidal currents, which helped to create a re-entrant feature. This boundary may be further evidence of sediment moving northwest, towards the nearshore zone between Milford and Hordle, which might provide sand for nearshore accretion. However, both Lacey (1985) and Nicholls (1985) identify a trend of net erosion immediately off Milford for the period 1880-1970, thus this sand may be moved westwards, via bar and trough topography, to the offshore pathway. A small-scale, and apparently closed, anticlockwise sand transport system is thus inferred.
Velegrakis and Collins (1994) report that over most of this area (as well as seaward) sand transport rates are low; sand may only be mobilised when peak tidal currents and waves with heights greater than 1m combine. Where water depths are less than about 9 m, slope gradients rarely exceed 0.5°, and are less than 0.1° close to the position of maximum low water (Halcrow, 1999). The inshore zone is demarcated seawards by a distinct break of slope, with water depths increasing to 20m immediately north of Dolphin Bank. Sediment transport in this deeper area has not been systematically researched. The recent swath bathymetry provides evidence that further offshore from Hordle, towards central Christchurch Bay, a separate set of bedforms are visible on the seabed. The bedforms run from east to west and the bedforms’ asymmetry indicates westward movement of sediment across the bay.
Two swath bathymetry surveys provide 100% coverage of Christchurch Bay. Through the Southeast Regional Coastal Monitoring Programme, the inshore 1 km zone of Christchurch Bay was completed in 2010. Further offshore, Poole and Christchurch Bays were surveyed in 2012 as part of the Maritime and Coastguard Agency UK Civil Hydrography Programme. A lobe to the north west corner of the Shingles Bank also follows the North Channel, parallel to Hurst Spit, and bedforms can be seen running in a north westerly direction along the bank and in the channel itself, perpendicular to the shoreline, confirming the transport of sediment from North Channel to Hordle (O2). The seabed nearshore between Mudeford and Milford-on-Sea and the seabed extending offshore in the centre of the bay is featureless, gently sloping seabed of gravel and sand, lacking bedforms. The nearshore seabed east of Milford-on-Sea becomes deeper and the sediment becomes coarser. This is due to the North Channel which runs between Hurst Spit and the Shingles Bank which creates increased current speeds in this area.
O3 Feed to Dolphin Bank
Offshore survey by echo-sounding, side-scan sonar and sediment sampling revealed sand megaripples to the south of Dolphin Bank which indicated northward transport onto and across the bank (Velegrakis, 1994), in addition to westwards movement, also determined by Dyer (1970). Southward sand transport is indicated by bedforms several kilometres to the north of Dolphin Bank which suggests that a sediment sink should exist in the central - east part of Christchurch Bay. However, sub-bottom profiling has revealed only a thin sediment cover and much bedrock exposure in this area. It is therefore concluded that sand is highly mobile in Christchurch Bay and transport pathways may vary seasonally in their relative strengths (Velegrakis, 1994; Brampton et al., 1998). Survey was completed during calm summer conditions and some of the bedforms identified may not be representative of long-term transport pathways, thereby explaining the partial disparity between inferred sediment transport directions and patterns of sediment accumulation. These indicated transport pathways are therefore of low to medium reliability and require verification. By contrast, survey in the vicinity of Shingles Bank (Velegrakis and Collins, 1993; Velegrakis, 1994) confirmed the transport pathways previously identified by Dyer (1970). This work indicated a capacity for long-term stability of the sediment transport pattern in this area (described in Section 4.1).
Dyer (1970) discerned, from the evidence of the asymmetry of sand waves resolved by echo-sounding, side-scan sonar and sediment sampling, the westward transport of progressively finer sands from the southern Needles Channel to Dolphin Bank. Velegrakis (1994), using similar survey techniques and applying sediment sorting co-efficients, concluded that there was a net westwards, and possibly south-westwards, pathway of sand movement over only the western and central areas of the Bank. Sediment movement along its northern flank, and at its eastwards extremity, was found to be moving north and north-eastwards. This implies a bedload transport parting close to the central crest. Velegrakis and Collins (1993) argue that superficial sand deposits that accumulate during the winter months over parts of the Shingles Bank derive from the eastern part of Dolphin Bank, where water depths are less than 8m. This material is mobilised and transported by waves, with the flood tide current a possible auxiliary mechanism. Once sand has moved across the crest line of The Shingles Bank, it is temporarily deposited on its eastern flank. The finer texture and better sorting of sand on this flank, compared to its western margin, is taken as a strong implication of a west to east transport pathway. To maintain this transport system, Velegrakis (1994) argued that there must be a supply of sand, moved by tidal currents, in the southern Needles Channel. Brampton et al. (1998) consider that as this source would supply only very fine sand, the composition of Dolphin Bank supports the probability of additional supply from the south. Both Nicholls (1985) and Lacey (1985) determined, from comparisons of hydrographic charts published between 1880 and 1968, that there was a net loss of sediment in the area immediately north of Dolphin Bank. This was estimated to have been approximately 60,000m³ per year over this period. Velegrakis (1994) carried out a similar analysis, covering 1979-1990 and identified a continuing erosional trend.
Near sea-bed current metering, undertaken during both the flood and ebb stages of neap and spring tidal cycles, indicated that tidal current velocities were sufficient to account for all observed movements of sand on, and marginal to, Dolphin Bank. This is confirmed by the modelling of tidal residuals from TELURAY and the application of critical flow velocity thresholds to that data (Brampton et al., 1998). The contribution of wave action to sand movement has not been determined, but is probable under storm wave conditions. This is especially likely where northward-moving sand is transported along the westward flank of The Shingles Bank, and its north-westerly salient.
Two swath bathymetry surveys provide 100% coverage of Christchurch Bay. Through the Southeast Regional Coastal Monitoring Programme, the inshore 1 km zone of Christchurch Bay was completed in 2010. Further offshore, Poole and Christchurch Bays were surveyed in 2012 as part of the Maritime and Coastguard Agency UK Civil Hydrography Programme. The Dolphin Bank extends westwards from the tip of the Shingles Bank and bedforms are visible along both flanks. It is difficult to confirm the movement of sediment around the bank based on this snapshot in time. Through bedform analysis it does appear that asymmetrical bedforms are present on both flanks. Bedforms on the southern flank indicate westward transport and bedforms on the northern flank indicate eastward movement of sediment. The seabed depth is greater than 15m therefore these are not likely to be wave driven bedforms.
O4 Transport from Dolphin Sand, and within the Central Bay
Some transfer from Dolphin Bank to Dolphin Sand is also considered likely, on the basis of analysis of bedforms in the intervening area (Brampton et al., 1998).
Two swath bathymetry surveys provide 100% coverage of Christchurch Bay. Through the Southeast Regional Coastal Monitoring Programme, the inshore 1km zone of Christchurch Bay was completed in 2010. Further offshore, Poole and Christchurch Bays were surveyed in 2012 as part of the Maritime and Coastguard Agency UK Civil Hydrography Programme. A variety of large and small bedforms extend along the seafloor westwards from the Dolphin Bank towards Dolphin Sands indicating continued sediment transport west (O4). These bedforms, however, do bypass Christchurch ledge and continue, uninterrupted, across Poole Bay to Old Harry Rocks (Handfast Point). Further offshore from Hordle, towards central Christchurch Bay, a separate set of bedforms are visible on the seabed. The bedforms run from east to west and their asymmetry indicates westward movement of sediment across the bay.
Summary
Survey and research analysis has been largely confined to the offshore area in the central and western parts of Christchurch, and in particular the Shingles Bank. There is, as yet, no conclusive evidence in favour of a confined, anticlockwise circulation pattern for bedload transport of either sand or gravel. Velegrakis (1994) and Brampton, et al., (1998) state that several discrete areas of small bedforms located on the seabed immediately east of Christchurch Ledge indicate a net north-east to south-west transport pathway. If this is a permanent feature, there can be no supply of sand back to this area from Dolphin Bank and Dolphin Sand. Hydrographic chart analysis indicates net erosion of the seabed of Christchurch Bay north of the Outer Banks, thus implying that sediment is leaving the system. In the nearshore zone, both net onshore and net offshore pathways have been recognised. These may be components of a sequence of small, anticlockwise-circulating systems of transport; Halcrow (1999) regard the onshore pathways as "sand bridges" that help to sustain the temporally variable but persistent pattern of nearshore shore parallel sand and fine gravel ridges or bars and intervening troughs. Tidal currents are considered the main mechanism of sand transport, but wave-induced stresses over shallower areas of the seabed are likely to promote onshore movement, especially during the winter period in relation to observed nearshore bar mobility.
Gravel transport is apparently confined to the Needles Channel and The Shingles Bank. Survey data and analysis of bathymetric change indicate both tidal and wave-induced mobility of gravel, with complex on and offshore patterns of movement. The Shingles Bank - and arguably the Dolphin Bank - constitute a large ebb delta established at the exit of the West Solent. It is less certain if the Needles Channel provides a significant pathway for gravel supply to Pot Bank.
4.2 Estuarine Output
EO1 Christchurch Harbour
Marine sediment input is possible at Christchurch Harbour entrance and a flood delta of well sorted sands and sandy gravels is located immediately inside the entrance (Murray, 1966; Tosswell, 1978; Gao, 1993; Gao and Collins, 1994a, b, c). This feature is attributed to transport and deposition on the flood tide (Murray, 1966; Gao, 1993), but detailed current metering at the entrance has revealed that ebb flow (peak surface velocity 1.9m per second) is significantly stronger than the corresponding flood (1.15m per second). Bedload transport is therefore expected to be predominantly out of the harbour, unless ebb and flood flow follow different channels, or flood flow coincides with storm waves to create intermittent sediment pulses into the harbour (Tosswell, 1978). Detailed sediment analysis involving sampling at 316 sites within the harbour revealed that sand is the dominant sediment type and that clay and silt content is generally low, except where marsh silts and river alluvium have accumulated. This indicated that suspended sediments tend to be moved out of the outer harbour (Gao and Collins, 1994a and b). However, analysis of foraminiferal tests (Gao and Collins, 1995a) indicated that those with an external (marine) provenance are predominant in sediments in the outer harbour, thus demonstrating that there is net accumulation here of biogenic debris introduced by tidal currents. This material compares closely with fine to medium sand in terms of transport potential. Thus, in this context, the tidal basin of Christchurch Harbour may be regarded as a net sediment sink, with additional sedimentation of fluvially derived material in the inner, especially the north-east area.
Current metering at the entrance has revealed that ebb tidal flow is significantly stronger than the corresponding flood (Tosswell, 1978). The tidal range is relatively small, 1.2m during spring tides, 0.7m during neaps. The tidal prism is 1.43M.m³, and the tidal regime has a slight double high water effect that accentuates the ebb current (ABPMer, 2009). Gao and Collins (1994b) give the mean ebb current speed to be 0.55m per second, that of the flood 0.165m per second. Maximum ebb velocities can attain 2.5m per second. This is because mean discharge from the rivers Stour and Avon is equivalent to the spring tidal prism, calculated as 1.06M.m³ (Gao and Collins, 1994, a; 1995, b). This would be enhanced under surge conditions. These hydraulic conditions indicate strong potential for bedload transport out of the harbour, but it is likely that actual output is small due to limited sediment supply to areas near the entrance where entrainment and seaward flushing are possible (Halcrow, 1999). It has been suggested that marine inputs under surge conditions may be significant, as indicated by a predominantly coarse sandy flood delta inside the harbour entrance (Gao and Collins, 1994a, c; 1995 b). Samples of exotic benthic foraminiferal tests taken from harbour sediments (i.e. derived from marine sources) indicate that their percentage in comparison to estuarine species is highest close to the entrance. Thus, marine-derived sediment (tests have similar mobility to sand grains) may not penetrate in quantity beyond the flood delta. The major effect of strong ebb currents is apparently to flush offshore any sediment transported longshore into the entrance channel from the opposing spits that confine the width of the mouth of the harbour. Thus, the channel is a littoral drift barrier, but with bypassing possible via the offshore bar (ebb delta) (Tosswell, 1978; Gao and Collins, 1994b). A relatively small quantity of fine suspended sediment is supplied by erosion of the harbour margins and seabed and by fluvial input (see Section 2.2). Despite this, the harbour sediments are predominantly sandy and reports suggest a trend for increasing sandiness (Gao 1993, Gao and Collins, 1994c). Hydraulic factors indicate that because the harbour has a high tidal exchange ratio there is a very thorough flushing effect on each tide (Tosswell, 1978; Gao, 1993; Gao and Collins, 1994a; 1995b). Qualitative observations indicate that sediments are readily resuspended by wave action inside the estuary and are efficiently discharged by ebb tidal flow (Hydraulics Research, 1987). An exceptional condition may occur when peak fluvial discharges, carrying suspended sediments, are impounded by coincidental high spring tides and/or a surge, allowing for some settling out. Incoming flood flow from Christchurch Bay is significantly less turbid (Gao, 1993). Gao and Collins (1994a) applied transport formulae to the relationship between mean current speeds through the harbour entrance; water level of the entire estuary basin and freshwater discharge. Bedload transport at the entrance operated at a much higher rate when tidal and wave-induced currents acted in combination. No net transport due to the flood tidal current alone was determined, but sediment discharge of a potential 8 x 10-²m³ per meter per second is considered possible when maximum ebb currents combine with high freshwater discharge. Thus, ebb transport is between one and two magnitudes greater than the flood, moving between 11,000 and 30,000m³ per year towards the ebb delta to a distance offshore of approximately 1km. This is net output - gross movement might be in the order of 100,000m³ per year.
The shallow entrance channel (mean width of 47m) has had a stable configuration - partly due to protection of Mudeford Quay - and cross-sectional area since at least the late 1930s (Gao and Collins, 1995b), thus indicating that it has probably achieved an equilibrium condition. Coarse sediment arriving at the entrance via littoral transport is, according to both observational evidence and mathematical modelling moved predominantly offshore (Gao and Collins, 1994a and c; 1995 a and b).
Christchurch Harbour, with a mean tidal basin area of 1.9 km² (reduced historically by reclamation and more recent landfill) and a mean water depth of 2.0m, is occupied by sandy gravels, sands, muddy sands and - towards its inner margins - silty muds. Average sediment thickness is 1.5m, increasing to over 2.0m in areas of bank accretion. There is a general consensus that an unknown proportion of fine sediments delivered by fluvial discharge are trapped within the estuary, and contribute to the prevailing vertical sedimentation rate of 0.1 to 0.4mm per year (Gao and Collins, 1994a; c). Fine sediment introduced by the flood tide may not, however, fully settle out, and a proportion is moved out of the basin on the ebb current (Halcrow, 1999). There is no quantitative evidence to support this assertion, but Gao and Collins (1995a, b, c) have indicated that spring tide ebb currents have substantial capacity to entrain fine to medium sand in the vicinity of the flood delta, leaving a gravelly infill in the lower main channels. Some internal re-distribution of sediment occurs in the vicinity of the northern coastline; this is apparent from analysis of sorting co-efficients and the evidence of small bedforms (Gao and Collins, 1994c). There is no reported evidence of significant erosion of mudflats, sandflats and saltmarsh margins, but abrasion of sub and inter-tidal flats may nonetheless be sources of suspended sediment loads. Maximum wave heights within the harbour vary between 0.4 and 0.8 m (1 in 10 and 1 in 100 year frequency, respectively, Hydraulics Research, 1987; Halcrow, 1999), due to limited fetch and may effect some marginal erosion. Harbour beaches are not well-developed, but the presence of gravel in places is probably derived from erosion of fluvial terrace deposits at a stage when the harbour entrance was more open to wave action, i.e. pre-dating the development of the present-day spits, see Section 5.2.
Both high and low multispecies saltmarsh are present, the latter colonised by Puccinellia and several other co-dominant grasses. Only small quantities of Spartina anglica are present, with no clear evidence of formerly more extensive occupancy. The sedimentation system of Christchurch Harbour has not, therefore, been significantly affected by the spread and subsequent "dieback" of this species, as experienced in most other south coast estuaries. Cope et al., (2012) were unable to provide any conclusive quantitative estimates of saltmarsh loss, 1971 to 2001, from air photo interpretation owing to imagery inaccuracies. There are, however, small areas of evident edge erosion. Any loss of saltmarsh seems to have been counterbalanced by its re-colonisation of abandoned artificial salt pans. Extensive Phragmites reed beds occupy tidal creek margins and areas of elevated marsh, where they front wet grazing meadows.
The overall sediment budget of Christchurch Harbour is positive, as evidenced by net vertical accretion. Small, but unquantified, inputs have derived in the past from the dredging of the lower channels of the Avon and Stour as part of flood alleviation schemes. Grimbury Marsh, for example, is substantially composed of spoil material. It is uncertain if this artificial source of sediment accretion continues at present. In overall terms, this estuarine system is a sediment sink.
EO2 Hurst Narrows
Ebb tidal currents are shorter lived, but more rapid, than corresponding flood currents at Hurst Narrows (Webber, 1980a), thus a dominant south-western transport pathway extends along the Needles Channel. Peak surface velocities of up to 3m per second have been recorded (Heathershaw and Langhorne, 1988) and up to 1.5m per second on the beach face at Hurst Point (Nicholls, 1985). Further offshore, peak surface and near bottom velocities of 2m per second and 1.2-1.4m per second respectively have been measured in the channel immediately east of Shingles Bank (Velegrakis and Collins, 1992; Velegrakis, 1994). Hurst Spit beach between the fossil recurves and Hurst Point slopes steeply seaward and is subject to rapid tidal current scour. These factors cause significant gravel removal, where material is readily entrained and transported seaward (Dyer, 1970; Nicholls, 1985; Velegrakis, 1994). This was calculated at 9,200-15,400m³ per year for 1939-68, based on map comparisons, and 8,700-11,000m³ per year for 1980-82, based on profile information (Nicholls, 1985). Surveys involving echo sounding, side-scan sonar and sediment sampling reveal the sea-floor at Hurst Narrows and westward along the Needles Channel to be composed of spreads of mobile gravel with intermittently present overlying sand and intervening patches of current swept bedrock (Dyer, 1970; Velegrakis and Collins, 1992; 1993; Velegrakis, 1994). Asymmetric gravel waves or megaripples have been identified throughout this area and all surveys have identified a consistent trend for south-westwards gravel transport along Hurst Narrows channel. Diminishing currents down this tidal transport pathway result in deposition of sediments, according to grain size. In this manner, some coarse material may be deposited on Pot Bank, but the majority is recirculated northward onto Shingles Bank (Dyer, 1970; Hydraulics Research, 1977; Velegrakis, 1994; Brampton, et al., 1998). Some sands are transported south of Pot Bank and deposited on a crescent-like feature, but the majority is transported westward towards Dolphin Bank and Dolphin Sand (Dyer, 1970). There is no indication of return northward transport from Pot Bank or the southerly sand features, so the limited evidence available suggests that these are sediment sinks (Hydraulics Research 1977, Brampton, et al., 1998). A more detailed examination of the evidence for transport pathways in this area is given in Section 4.1.
Two swath bathymetry surveys provide 100% coverage of Christchurch Bay. Through the Southeast Regional Coastal Monitoring Programme, the inshore 1 km zone of Christchurch Bay was completed in 2010. Further offshore, Poole and Christchurch Bays were surveyed in 2012 as part of the Maritime and Coastguard Agency UK Civil Hydrography Programme. The swath bathymetry within Christchurch Bay highlights the complexity of the offshore sediment transport through the Shingles and Dolphin Banks. Starting at the Shingles Bank, which extends southwest from Hurst Spit, bedforms can be seen running along its entire length. The bedforms indicate a lateral flow of sediment southwest along the bar. In the channel, bedforms can be identified in the same orientation, running from northeast to southwest, along the channel illustrating sediment leaving the West Solent (EO2).
EO3 Input to West Solent
Suspended sediment analysis using Landsat imagery (MacFarlane, 1984) revealed high concentrations in Christchurch Bay, with a marked peak in the winter months (Lacey, 1985). This implied cliff erosion to be the major source because it too shows similar seasonal variation. Examination of suspended sediments in the Western Solent using remote sensing also yielded a pattern of peak winter concentrations and indicated significant transfer of suspended sediments from Christchurch Bay to the West Solent (Srisoengthong, 1982; Lacey, 1985). Net suspended transport is likely to be into the West Solent at Hurst Narrows due to the greater duration of the flood current (Webber, 1980a). Sediments transported into the West Solent in this manner are deposited on the north-west shore of the West Solent and therefore represent an output from the Christchurch Bay system. Small quantities may be returned to Christchurch Bay because Hurst Spit is receding over the Keyhaven marshes and approximately 9,000m³ per year of fine sediment is released by erosion of the seaward face (Nicholls, 1985). However, the precise fate of this material is not known (Posford Duvivier, 1999). HR Wallingford (1994) undertook a brief programme of sampling of suspended sediments off the distal point of Hurst Spit during spring tides. Concentrations varied from 0.2 to 30 mg/l, with a mean value of 5.0 mg/l. The suspended flux of silt and clay over a single tidal cycle may be 2 to 15,000kg per second, but how much is lost to either the Western Solent or offshore is not known.
4.3 Wave Driven Offshore Loss
Hengistbury Head
Sea-bed drifter studies, tracer experiments and physical model studies all indicate net offshore transport at Hengistbury Head (Watson, 1975; Tyhurst, 1976; Wright, 1976; Webber, 1980b; Halcrow, 1980; Wright, 1982; Hydraulics Research, 1986a). Physical model studies showed that offshore transport was predominantly of sand in the vicinity of the Long Groyne. The model covered a limited area and no return shoreward feed was recognised (Hydraulics Research, 1986a). This may indicate output to a sediment store or sink to the south of Christchurch Ledge, although the consensus view has been that sediment is swept north-west off the Ledge, where it may contribute to the beaches of Mudeford Spit and Highcliffe via the sandbank system of the ebb tidal delta seaward of the entrance to Christchurch Harbour. This information was used to inform the 2004 WO1 arrow, to indicate wave driven onshore to offshore sediment transport on the eastern side of Long Groyne, with transport occurring to the south-east over Christchurch Ledge. The supporting text is contradictory and therefore the arrow has been removed from the 2012 update.
Barton-on-Sea
The previous SCOPAC (2004) STS suggested that there was wave driven onshore to offshore transport of sediment in the vicinity of Barton-on-Sea with low reliability (WO2). This was based on the literature available but does not reflect processes occurring after the installation of rock protection works completed in 1990. Analysis of Coastal Monitoring Programme multibeam bathymetry data (2010) covering the nearshore 1km zone in Christchurch Bay reveals a featureless, gently sloping seabed of gravel and sand, lacking bedforms. This therefore does not support an onshore to offshore pathway at Barton-on-Sea. The WO2 arrow has therefore been removed from the 2012 STS update.
Examination of beach profile data covering the period 1974-82 yielded erosion of the intertidal zone of West Barton by 1,000m³ per year and the nearshore zone, to -5mOD, by 19,000m³ per year (Lacey, 1985). This coincided with erosion of the offshore zone by 21,000m³ per year and lowering of 2m determined from bathymetric chart comparisons for the period 1880-1968 (Lacey, 1985). Substantial offshore loss was also predicted by wave energy flux analysis, which indicated 165,000m³ per year (Henderson, 1979) and 296,000m³ per year (Lacey, 1985). This analysis may be subject to error, thus the reliability of an offshore pathway is uncertain; intensive profiling covering the period 1976-77 revealed net beach accretion and significant short-term variability due to slow onshore bar movement in calm conditions and its rapid destruction and offshore transport during storms (Babbedge, 1976a; Webber, 1980b; Halcrow, 1980).
Hordle
Beach profiling over the period 1981-82 revealed a loss of 23,500m³. Seasonal morphodynamic variation was significant, with offshore loss during storms and onshore input during calm conditions (Nicholls, 1985). The long-term trend determined from hydrographic chart comparisons was for net accretion of 2,600-2,900m³ per year for 1867-1939, rising to 8,400m³ per year for 1939-68 but then declining to 5,900m³ per year for 1968-78 (Nicholls, 1985). Velegrakis (1994) also identified net accretion, 1979-1990, from chart analysis, although quantities are not given. Several possible conclusions may be drawn; (i) profile data was collected over a limited time period and was unrepresentative of long-term trends; (ii) chart analysis may reflect an underlying trend for accretion at Hordle; (iii) profile data extended to -1mOD, further seaward than the MHW line examined by chart comparisons, thus the erosion revealed may reflect sediment loss restricted to the lower foreshore and nearshore zones. Comparison of charts covering the period 1880-1968 revealed sea-bed lowering up to 3m off Milford which supports this latter possibility (Nicholls, 1985; Lacey, 1985). Offshore survey by echo sounder, side-scan sonar and sediment sampling revealed asymmetric sand megaripples which clearly indicated offshore transport of sand (Velegrakis, 1994). The fate of this material is uncertain because although bedforms indicated net transport towards central-east Christchurch Bay sub-bottom profiling showed that sediment cover here was thin (Velegrakis 1994). A corresponding pathway was not identified by Dyer (1970). Compilation of a beach sediment budget for the eastern part of Christchurch Bay and detailed beach sediment sampling revealed progressive eastward offshore sand loss (Nicholls, 1985).
The previous SCOPAC (2004) STS suggested that there was a wave driven onshore to offshore sediment transport pathway in the vicinity of Hordle (WO3), with low reliability. This was based on the literature available however does not reflect processes after the installation of rock protection works west at Barton-on-Sea completed in 1990. Analysis of Coastal Monitoring Programme multibeam bathymetry data (2010) covering the nearshore 1km zone in Christchurch Bay reveals a featureless, gently sloping seabed of gravel and sand, lacking bedforms. This therefore does not support an onshore to offshore pathway at Hordle. The text above refers to sediment movement from onshore to nearshore transport of fine sediment. The WO3 arrows which were located at Hordle have therefore been removed from the 2012 STS update.
Overall Trends
Sediment eroded from beaches in Christchurch Bay did not balance with inputs from littoral drift, onshore transport and cliff erosion, so was a presumed net loss of beach sediment offshore (Lacey, 1985). The majority of sea bed sediments in Christchurch Bay are thus derived from long-term cliff erosion. Clays are transported out of Christchurch Bay in suspension (Lacey, 1985), whereas sand is transported as bedload offshore and deposited in the outer Bay (Dyer, 1970; Nicholls, 1985; Velegrakis, 1994). Gravel is retained on the beach, but is transported eastward by littoral drift and is subject to some abrasional wear. Coarse material reaching the distal zone of Hurst Spit is transported offshore by rapid tidal currents at Hurst Narrows, and into the Needles Channel. Hydrographic chart comparisons of the whole of Christchurch Bay revealed an equivalent net loss of 505,000m³ per year over the period 1868-1968 (Lacey, 1985). If this is accurate, it indicates that it is a site of net output, with a continuing tendency for selective offshore loss of beach sediment. It is therefore not a fully self-contained or confined transport cell.
5. Sediment Stores and Sinks
Specific details of beach morphodynamics, most of them relating to short-term and relatively recent surveys and analysis are given in the preceding section on littoral transport and are also detailed in Webber (1980b) and Nicholls and Webber, (1988a and b). The first section below provides a longer-term overview of beach behaviour, adding further information where available. The second section addresses the evolution of the spits enclosing Christchurch Harbour, and Hurst Spit, together with a summary of their contemporary morphological, morphodynamic and sedimentological characteristics. Further information may be found at http://www.scopac.org.uk/sediment-sinks.html.
5.1 Restrained Beaches: Highcliffe to Milford-on-Sea
A. Highcliffe
Groyne construction commenced in the mid-nineteenth century and for the next 80 years these structures were effective in retaining a stable beach. However, the more influential factor was the rapid growth of Mudeford Spit during this period (see section 4.2). This provided a sediment source and significantly reduced nearshore wave energy. At its maximum extent, the distal end of the spit was nearly co-incident with the position of Highcliffe Castle estate (Burton, 1931; Robinson, 1955). Repetitive breaching and a tendency to landward migration caused substantial beach accretion, particularly in the 1920s and early 1930s, at Avon Beach. Beach volume loss commenced in the early 1940s, and continued through to the late 1970s, after which a coordinated programme of beach nourishment and groyne construction was initiated. Nonetheless, there was a 21% loss of volume between 1985 and 1990 (Hydraulics Research, 1991; Tyhurst, in Bray and Hooke, 1998) but modest losses thereafter. The submerged sand bar identified by Gao and Collins (1994-1995), which is further seaward of the breaker zone by comparison to similar examples in Poole Bay, is of uncertain morphodynamic status. It may indicate net onshore transport, but could be a partial product of earlier renourishments. Alternatively, it might represent in part the foundations of the former Mudeford Spit, which is now effectively a component of the complex ebb delta seaward of the mouth of Christchurch Harbour.
Throughout the 1990s, net beach accretion was promoted by further limited renourishment and the sequence of robust rock groynes that extend to Chewton Bunny. The underlying trend, however, is erosional. The terminal groyne promotes local wave diffraction, and thus immediate updrift beach retention. Gao and Collins (1994-1995) infer, from limited bathymetric analysis, that there is net offshore (seaward) sediment movement, and thus potential beach drawdown, in the main rock groyne sector.
B. Naish Beach
The mixed sand and gravel beach fronting the eroding cliffs of this sector has experienced nearly continuous volume loss, and recession of the position of mean low water, since the late 1960s. One cause, at the western end, is the angled terminal rock groyne that flanks Chewton Bunny, and which sets up wave diffraction (Mackintosh and Rainbow, 1996). Beach monitoring since 1988 indicates an average annual depletion of approximately 4,000m³ (Halcrow, 1999). In an attempt to address this problem, without recourse to cliff stabilisation, a proposal to combine renourishment with a set of "dynamic groynes" built of shingle was put forward. These would promote wave energy dissipation whilst also providing a downdrift sediment supply if and when eroded. Physical modelling, using a groyne spacing of 200m demonstrated that these structures would dynamically evolve, although their efficiency might depend on periodic sand recycling and/or gravel recharge (HR Wallingford, 1995). A clear advantage of this innovative approach would be the maintenance of sediment input from cliff erosion, and the bridging of the "pinch point" between the adjacent defended frontages (Mackintosh and Rainbow, 1996; Bray et al., 1996). To date, this scheme has not been implemented.
C. Central and East Barton (east to Becton Bunny outfall)
Hooke and Riley (1987) and Halcrow (1999) have analysed serial Ordnance Survey maps and air photos, and identified the retreat of both MHW and MLW between 1867 and 1969. MLW retreat over this sector was almost twice as fast as that of MHW (approximately 0.80m per year), thus leading to beach narrowing and steepening. With the fixing of the position of MHW following the installation of a basal rock revetment along the central Barton shoreline, this trend has continued since the late 1960s.
D. Becton Bunny to Hordle
Immediately eastwards of Becton Bunny outfall there has been a marked tendency towards beach steepening. Between Taddiford Gap and Hordle (Rock Cliff), both MHW and MLW moved seawards, at 0.34m per year between 1870 and 1970 (Hooke and Riley, 1987). This has been ascribed to a persistent tendency for onshore bar migration and the accretion of a small foreland feature (Gao and Collins, 1994-5). Halcrow (1999) identify net beach width reduction (but no apparent steepening) commencing in the early 1970s, at rates of up to 2m per year of LWM recession. The reasons for this switch from accretion to erosion is not clear, but is likely to be linked to some loss of onshore feed.
E. Milford-on-Sea
MLW retreat at 0.65m per year, and MHW recession at 0.45m per year occurred between 1867 and 1969 (Hooke and Riley, 1987). This trend towards steepening has continued since the installation of the sea wall in 1970 and subsequent revetment (Webber, 1980b; Halcrow, 1999). The gravel upper beach face has an average gradient of 1:6, whilst that of the more sand dominated inter-tidal beach varies between 1:10 and 1:20. The adjacent sub-tidal profile is characterised by up to three shore-parallel segmented sand and sandy-gravel bars that are spatially and temporally variable. (Karunarathna et al., 2012 provide details of dimensions, sediment properties and morphodynamic character, the latter based on a continuous 18 year record of repetitively measured beach cross-shore profiles). Overall volume loss has occurred during recent decades, partly as an outcome of the natural reflective behavioural character, augmented by sediment yield reduction effects of the seawall; and partly because of winter storm erosion of a nearshore sand bar (Halcrow, 1999). This has occurred despite the presence over several decades of timber and, more recently, rock groynes. There has been a long-term erosion “hot spot” immediately downdrift of the revetment protecting the White House due to attempted beach realignment by longshore transport in response to wave climate. In addition, net offshore transport feeds an impersistent sub-tidal shore-parallel gravel bar, characteristically during storms. Generally, beach levels fell approximately 2m in the ten years following renourishment in 1985 (Mackintosh and Rainbow, 1995). Mortlock (2012) has shown that run-up elevations increase when the profile of Milford beach (and other regional examples) is modified by bimodal wave spectra (storm and sea waves combined). As a result of partial collapse of the promenade here in 2008, a rock revetment was constructed in front of the White House and 5,000m³ of gravel placed on top. However, continuing reduction of levels and volume along the length of this swash dominated beach has necessitated recharges, each of 5,000m³ in 2010, 2011 and 2012 together with crest elevation and slope reprofiling (Channel Coastal Observatory, 2012). Horrillo and Reeve (2010) analysed 29 beach profiles repeatedly surveyed between 1988 and 2004, together with data on transformed nearshore significant wave heights derived from numerical hindcasting and-from 2003 a wave rider buoy located offshore Milford-on-Sea. They conclude that 85% of profile variability could be ascribed to temporal variability of wave height and period. The continuing presence of a shore parallel sand bar, and its effect on incident wave steepness, was apparent. Using essentially the same data sources, but different statistical analysis, Karunarathna et al., (2011) argue that overall profile response was most directly conditioned by incident wave steepness.
General
All researchers are agreed that the beaches of Christchurch Bay are currently declining in volume, and have been doing so for several decades. The annual loss is between 10-20,000m³ (Webber, 1980b; Nicholls and Webber, 1988a and b; Nicholls, 1985; Halcrow, 1999), a range that reflects inter-annual spatial and temporal variability of forcing conditions.
The relatively small tidal range concentrates wave energy into a narrow beach zone. The nature of the wave climate determines the fact that the highest rates of volume loss, and morphodynamic change, have been along the central-eastern sector. Wave energy is less along the far eastern and western sectors, with relatively less change (see sections 5.2 and 5.3).
Gao and Collins (1994-5) conclude that, although the overall trend for the past 80-100 years (at least) has been one of beach narrowing and steepening, it is not a consistent recent feature of all beaches fronting Christchurch Bay. However, this conclusion was based on the analysis of a relatively short temporal sequence of beach profiles.
The latter (a total of 44) also revealed the presence - especially in winter - of shore-parallel bars or ridges, extending into water depths of between -5 and -10mCD (Gao and Collins, 1994; 1995). Most were presumed to be dominantly composed of sand (or sandy gravel in a few cases), and to indicate seasonal profile change rather than any steady net onshore transport in the nearshore zone. There is some, ambiguous, evidence that the persistence of these features has been declining over the past 20-25 years, especially along the eastern sector.
Although median grain size coarsens eastwards, as a function of wave climate and selective removal of the fines fraction, beach gradients are not precisely adjusted to this control variable.
5.2 Mudeford Spits
The entrance channel to Christchurch Harbour is confined by spits of unequal length. To the south, attached at its proximal end to Hengistbury Head is Mudeford Sandbank, with an approximate south-west to north-east orientation. To the north is the smaller, but wider, Mudeford Quay spit which projects into the harbour on an east-north-east to west-south-west orientation.
The above geographical terminology is confusing, and for the purposes of this account Mudeford Sandbank will be referred to as Mudeford Spit, whilst the opposing structure will be termed Haven House spit. Through the analysis of estate plans, hydrographic charts, topographical maps and other archival sources, the evolution of both spits over the past three hundred and fifty years has been reconstructed (Burton, 1931; Robinson, 1955; Lacey, 1985, Bray et al., 1996, Christchurch Borough Council, 1999; HR Wallingford, 1999b; Halcrow, 1999). Mudeford spit has been especially dynamic, although interpretation of the evidence of change is made more complex by the nearshore presence of the ebb tide delta of Christchurch Harbour (Gao and Collins, 1994a and b; 1995a, b and c). Full morphodynamic understanding of these accretion forms remains elusive (Halcrow, 1999).
5.2.1 Mudeford Spit
Mudeford Spit currently extends approximately 900m in length, although historically it has been much longer and evidence shows it has been a dynamic and mobile feature (Royal Haskoning, 2011). The basement sediments are mixed gravel and sand, overtopped by fine sand that has generated small dunes and sandhills up to 7m in height. Earliest evidence for the presence of this spit dates to 1660, when an abortive attempt was made to construct an artificial channel through its narrowest point. This had infilled with sediment within 50 to 60 years, although Clarendon Rocks (extending some 250m seawards) remain as a legacy of this intervention. Between 1760 and 1785, the spit was approximately 8,900m in length, with the harbour entrance channel more or less on its modern alignment. Rapid extension occurred between about 1850 and 1880, involving some 2,500m of shore-parallel growth. The harbour exit in 1880 was opposite Highcliffe Castle estate, with the spit enclosing a 3,000m long narrow channel, The Run, between the old and new entrances to Christchurch Harbour. The reason for this dramatic change was anthropogenic: ironstone mining of Hengistbury Head between 1847 and 1856 caused a very rapid increase in cliff recession, providing a large quantity of sand for north-eastwards longshore transport. In addition, the by-passing of Hengistbury Head of sediment moving eastwards from Poole Bay was greatly accelerated. Much of this "surge" of sediment supply was used to build the extension of Mudeford spit in the 30 years following the cessation of opencast mining. It also increased in width over this period, from an average of 50m in 1840 to 300m in 1885, and acquired substantial crestal dunes. Although no further lateral growth occurred after 1880, Mudeford spit remained a substantial feature up to the 1930s. Nonetheless, it was temporarily breached under storm conditions several times, at a mean frequency of once every twelve years (e.g. 1895/6; 1911; 1924 and 1935). At the same time, it "rolled back" along segments between breach sites, to "weld" with Avon beach (Highcliffe) behind.
Formal protection started in 1931, concentrating on low points close to the distal point. Following the construction of the Long Groyne, at Hengistbury Head, in 1938 there was near instantaneous reduction in updrift longshore sediment supply. In response, the spit had reduced to a length of 900m by the mid-1940s, and had acquired its present configuration by 1948. The breach of 1935 was converted to a new harbour entrance channel, thus cutting off the downdrift distal sector; this subsequently transgressed landwards, to form a wide beach that temporarily trapped a lagoon. The site of the extended spit, as it was prior to 1935, continued to be marked by a wide, submerged sandbank; this feature persists to the present time, confining the channel of The Run, and constitutes the major part of the ebb delta constructed by tidal current outflow from Christchurch Harbour.
Since the late 1940s, protection works have stabilised the basic morphology of the spit. Nonetheless, the submerged sandbank beyond its distal point expanded, in stages, up to 1973. Since then it has shown the reverse trend. This apparent cycle of accretion followed by erosion may have occurred during the phase of spit extension, with a periodicity of 15-20 years. Rapid erosion of the distal tip has occurred under high energy south or south-easterly incident waves, but is now restrained by a rock revetment and sheet steel sheathing.
A progressive programme of installing concrete and rock groynes, and, more recently, sand renourishment has been carried out in stages between 1945 and 1996, together with restoration of longer-term recreation-induced erosion of the dunes (now only a small remnant of the dune field present in the late nineteenth century). Renourishment in the early 1990s, maintained subsequently by annual recycling, has maintained beach stability, which also benefited from a modest increase in sand supply following the recharges of Bournemouth beaches in 1974/75 and 1988/89. (The dynamics of sand by passing of the Long Groyne are discussed in detail in the chapter covering Poole Bay, and also in part 3 of this section). A net advance of MHW within groyne bays immediately downdrift of the Long Groyne has been the trend between the mid-1900’s and 2008 suggesting either a direct sediment feed from offshore or enhanced updrift supply (Royal Haskoning, 2011). Modelling of beach behaviour (HR Wallingford, 1999b), in the context of proposals to both widen and heighten Mudeford Spit (Christchurch Borough Council, 1999) indicate a 70% probability of breaching under wave energy conditions with a 1 in 100 year return frequency.
5.2.2 Haven House Spit
The natural form of this spit has been substantially modified by protection structures, especially the seawall/promenade built following the breach threat in 1950. It has an expanded distal head, connected to Mudeford beach by a narrow (30m) proximal "tail". Haven House was built around 1700, so the feature dates at least to his time. Evidence suggests it has a gravelly sand foundation, with overlying sand that was formerly a small set of dunes or sand hillocks.
In relation to Mudeford Spit, it is not only considerably smaller, but is offset to it - it is a type of apposition spit, with vestigial evidence of a distal recurve. Brampton et al., (1998) suggest that it might have been constructed as a conventional spit by littoral drift moving towards the west, supplied by cliff erosion at Highcliffe and further east. This introduces the necessity for localised drift reversal, for which there is no contemporary or recent evidence. Brampton et al. (1998) also suggest the possibility of "swash-bar welding", presumably similar to foreland construction by shore-normal waves. This might have occurred prior to any phase of extension of Mudeford spit, and would be characterised by several sub-parallel gravel and/or sand ridges - now concealed, and possibly destroyed, by development. Robinson (1955) preferred to interpret Haven House spit as a truncated sector of an ancestral (pre-1600) Mudeford spit that extended across the embayed mouth of the Avon and Stour rivers, i.e. pre-dating the modern form of Christchurch Harbour. It was subsequently permanently breached either by a major storm surge, or by hydraulic pressure set up by exceptionally high river discharge. Indeed, both conditions might have operated simultaneously.
Robinson's model was applied also to the entrances to Poole and Pagham Harbours, where more recent research has revealed alternative explanations. However, in the case of the "twin" offset spits of Christchurch Harbour, it continues to be an attractive hypothesis. The more appropriate context is that of mid to late Holocene landward barrier migration, for which this part of Christchurch Bay might be a suitable setting in terms of hydrodynamics and sediment supply. Given that a formerly much larger Hengistbury Head salient would have been an impediment to sediment supply from Poole Bay, onshore-directed transport is a plausible alternative. Further research on the sediment stratigraphy of Christchurch Harbour, and of the spits themselves, is needed to resolve this problem. The "roll over" behaviour of segments of the extended Mudeford spit between 1880 and 1935, and its subsequent "welding" onto Avon beach, is consistent with barrier morphodynamics.
5.3 Hurst Spit - Introduction
This internationally known multi-recurved barrier spit has a main axis approximately 2km in length, orientated at 130°N, whose proximal end is attached to the shoreline at Milford-on-Sea. At Hurst Point, its 800m distal sector recurves very sharply to a north/north-west orientation of 100°N, and terminates with an active recurve aligned westwards. Three former, now inactive recurves are present to the west of the modern feature, increasing in size and morphological complexity from west to east. Prior to its substantial modification during the 1980s, particularly along its proximal and central sectors due to a series of major storm surges, it had a crest elevation of between 2 and 4m above mean sea-level. This has been raised to 7m along its proximal sector, tapering to 5m at Hurst Point as a result of a comprehensive programme of stabilisation completed in 1996. The overall planform of Hurst Spit appears to have been stable since at least the mid-eighteenth century; Hurst Castle, close to the point of distal recurvature, was built in 1544 and has apparently only been subject to potential threat from beach recession over the most recent 50 to 80 years. Nonetheless, the spit is a dynamic landform that has adjusted to the impacts of historically infrequent major storms by steadily receding landwards. Its behaviour as a barrier structure has been evaluated fully in recent years (Nicholls, 1984 and 1985; Bradbury, 1998 and 2000; Bradbury et al., 2005, 2009).
5.3.1 Geomorphological development
As Christchurch Bay was opened out during the mid-Holocene due to sea-level transgression (Section 1), dominant south-westerly waves drove sediment both onshore and alongshore in a west to east direction. A substantial proportion was gravel, derived from the erosion of Pleistocene river terrace deposits originally deposited by the Solent River (West, 1980; Nicholls, 1987). An ancestral form of Hurst Spit developed following the creation of the entrance to the Western Solent circa 7000 to 6,500 years BP (Nicholls and Webber, 1987a; Nicholls, 1987; Velegrakis, et al., 1999). With continuing sea-level rise both updrift cliff recession and offshore sea bed erosion released large quantities of coarse sediment that created the Shingles Bank. This provided a large store of material which, together with a rate of littoral drift up to seven times what it is at present (Nicholls, 1985) created a substantial barrier spit. Low wave energy conditions to its lee promoted mudflat and saltmarsh accretion. Nicholls and Clarke (1986) have described a truncated sequence of estuarine muds and peat deposits that outcrops seawards of the modern beach face, indicating that spit recession was in progress at least by approximately 4,500 years BP. This process is presumed to have been continuous (in a time-averaged sense) since then.
The progressive south-eastwards growth of Hurst Spit appears to have been episodic, or phased, as indicated by the three main "fossil" recurves. None have yet been precisely dated, but each must represent a stage of temporary equilibrium between sediment supply and loss. As now, loss of sediment at the distal end would have been due to a combination of wave action and tidal currents; waves propagating into the Western Solent, and refracted by the Shingles Bank, would have caused distal curvature and powered south to north littoral drift. Tidal currents would have increased in velocity and capacity to transport sand and fine to medium gravel, due to narrowing of the entrance channel between the Isle of Wight coastline and the Spit terminus. The dominant pathway of gravel transport then, as now, would have been towards the Shingles Bank, thereby establishing a sediment circulation that sustained growth. (King and McCullough, 1971; Nicholls and Webber, 1987a; Velegrakis, 1994). The ultimate position of the distal point - Hurst Point - may not have been achieved until the historical period and it was determined by the presence of a steep, tidally-scoured slope at the beach toe that prevented any further accretion. This stability is evident form the fact that the modern distal recurve is substantially larger than its predecessors, and that the latter increase in size with decreasing age. There may, however, be additional explanations for earlier phases of distal recurvature, such as short-term sea-level still-stands; "pulses" of gravel supply from submerged sources or differences in wave climate. Alternatively, major storms might have caused barrier breakdown over the proximal sector, introducing large quantities of sediment into both the longshore and onshore transport pathways feeding the distal end. This would imply that forcing conditions and morphodynamic response were similar to those prevailing today (Bradbury, 1998).
5.3.2 Hydrodynamic regime
Hurst Spit has three distinct sectors of alignment (Lewis, 1938; King and McCullough, 1971, Bradbury, 1998), each of which is in response to spatial variation in wave climate. This is due to complex interrelations between wave approach, nearshore and offshore bathymetry and wave energy that declines from west to east (Hydraulics Research, 1989a; HR Wallingford, 1992; 1993; Wimpey, 1994; Bradbury, 1998). Details are set out in section 1, in the context of Christchurch Bay as a whole. The presence of the Shingles Bank causes shoaling, refracting and diffracting effects, which - together with the proximity of the Isle of Wight - results in mean significant wave height being one metre lower along the sector 700m west of Hurst Point in comparison to the proximal segment.
Tidal current velocities, both ebb and flood, are high, especially along the distal sector where tidal flow is constricted at Hurst Narrows. Here, where ebb current speeds can exceed 2.5m per second, significant quantities of coarse sediment are transported away from the spit and are moved offshore. Full details are given in sections 1 and 4.1.
5.3.3. Littoral sediment transport
As stated in Section 3, there is uncertainty concerning contemporary rates of throughput by longshore transport. Because each of the three main segments are swash-aligned, rates are less than they are on confined beaches further west. The rock armour over the 800 m sector between the neck of Hurst Spit and the detached breakwater installed in 1996 has reduced the natural rate of drift, which accelerates immediately downdrift. An average rate of drift between 11 and 15,000m³ per year along the central, unprotected, sector has been determined from various studies (Nicholls, 1985; Halcrow, 1999). As well as input from updrift beaches, and probably from offshore, a proportion of throughput derives from beach erosion. Using evidence of recession rates (see Section 4.3.5); and borehole data for the average thickness of gravel beneath the main axis of the spit and relict recurves, Nicholls (1985) calculated a potential supply of 1,600-3,900m³ per year for 1979-1982. This has diminished from 8,500-9,700m³ per year (1938-1968) and 1,600-5,600m³ per year (1969-1978) due to protection measures introduced in 1968/9 and a steady steepening of the beach face. Several episodes of severe beach erosion during the 1980s and early 1990s would have increased annual rates, which will also have been enhanced by several renourishments of limited quantities previous to the introduction of nearly 300,000m³ in 1996, doubling the previous volume. Longshore transport rates between 1998 and 2003 were approximately 45% (9,000m³ per year) of those predicted by design modelling (16,000m³ per year) and declined further to circa 5,000m³ per year between 2004 and 2008 (Bradbury et al., 2009). Inter-annual variability of +/-4,000m³ per year occurred in response to changes in incident wave approach directions, with evident spatial variation due to longshore differences of wave energy. (An example of this is provided by the slight ness feature that has developed downdrift of the headland breakwater). These lower rates of littoral drift may be ascribed to the beach becoming progressively more swash aligned (Bradbury et al., 2009). Virtually no accretion has been recorded immediately updrift (i.e. west) of the breakwater, thus suggesting that this original, pre-recharge, erosion prone location was due to a low rate of supply by longshore transport.
Between Hurst Point and North Point, the mean longshore drift rate is lower than along the more exposed westerly sector. Most of the sediment moved is coarse sand and gravel. It has been estimated that approximately 10,000m³ per year is removed into Hurst Narrows due to wave focusing caused by a deep scour hole off Hurst Point (Nicholls, 1985). This reduces the downdrift transport rate, as does the reduction in wave height along this northward-orientated sector. Prevailing rates may be between 3-5,000m³ per year (New Forest District Council, 1992; 1996), a quantity apparently dominated by fine gravel.
5.3.4 Sediment composition
Most information is restricted to sampling of clasts from the beach face and crest (Bradbury, 1998). The dominant constituent is sub-angular to sub-rounded flint gravel, with a mean diameter of 15 mm. Median clast size diminishes slightly from west to east in the direction of wave energy reduction (Nicholls, 1985). Contemporary sediment character has been modified by several replenishments in the 1980s using gravel from inland sources that had different size and shape characteristics than the indigenous material. The 1996 stabilisation scheme introduced some 280,000m³ of gravel dredged from the Shingles Bank. As this is a store for sediment moving away from, and then back to, Hurst Spit, its textural indices are similar to the "natural" population. Limited evidence (Bradbury, 1998; Bradbury et al., 2003, 2009) suggests that the proportion of sand to gravel increases slightly with depth, but does not prevent infiltration and cross-barrier percolation (Nicholls, 1985). Both cross-shore and longshore grading are relatively poorly developed, the reasons for which have not been specifically investigated.
5.3.5 Morphodynamic behaviour: Historical trends to 1980
The main processes that have controlled the cross-profile form, and steady landward recession, over this period are berm formation, over-washing and overtopping (Bradbury and Powell, 1992; Bradbury, 1998; 2000). The last two are associated with surge conditions and other factors creating high water levels and high energy breaking waves at or above crest level. Overwashing involves swash passing over the crest and then running down the back slope towards the back barrier saltmarsh (and incised channels). Crestal breaching occurs under conditions of overwash sluicing, which creates low crestal points that facilitate further overwash. It is almost certain that there have been numerous overtopping and overwashing events over the past 3-4,000 years of the history of Hurst Spit, but the first to be fully documented occurred in 1954. Analysis of successive editions of large-scale maps and plans; air photographs and hydrographic charts reveals that these processes, including crest cut-back, have resulted in net landward recession by rollover since the mid-nineteenth century. For the sector between Milford-on-Sea and Hurst Point, the rate was 1.0-1.5m per year, 1867-1968 (May, 1966; Halcrow, 1982; Nicholls, 1985; Hooke and Riley, 1987). It is probable that recession occurred intermittently during this period, with phases of stability (e.g. 1890-1910) alternating with short-term retreat of several meters under occasional high magnitude wave conditions. Spatial variations in the mean retreat rate can be deduced from the historical record, e.g. maximum recession of 1.8m per year occurred at the "neck" of the spit between 1931 and 1965, with a minimum retreat of only 0.2m per year over the central corridor during the same period (Hooke and Riley, 1987). Retreat rates everywhere appear to have accelerated (though variably) after about 1940, and Nicholls (1985) concluded that the spit as a whole moved landwards at 3.5m per year, 1968-1980. If no major overwashing events occurred during the period before 1950, most of the displaced material may have moved offshore. This would have resulted in progressive beach narrowing and steepening, a response that is implied from plotting the movement of the positions of mean low and mean high water (Hooke and Riley, 1987). Steady volume loss is therefore a feature of Hurst Spit over at least the past 140 years; from a programme of direct field measurements in 1980-1982, Nicholls (1985) calculated this to be 14,000m³ per year (but only 7-8,000m³, 1970-1979).
Accelerating recession in the 1950s and 1960s posed particular protection problems at the far proximal end, where breaching might - potentially - result in spit detachment. Thus, in 1969, massive rock armour was placed along a 600m frontage. Although effective as a protection measure, it reduced the rate of longshore drift and provoked erosional outflanking (Mackintosh and Rainbow, 1995) and the creation of a downdrift erosion “hotspot”.
The position of Mean High Water along the distal sector also moved landwards during the period 1870 to 1980, at a rate of 0.9m per year (Halcrow, 1999). Volume loss at Hurst Beach was calculated by Nicholls (1985) to have been 1-2,000m³ per year, 1965-1982. However, North Point, at the recurve tip, advanced some 60 m during this period; it deflected the mouth of the Keyhaven River across adjacent saltmarsh and mudflats, causing erosion. Previous to about 1920, possibly as far back as the mid-seventeenth century, North Point was the site of aggregate removal. As it was more or less stable during the period 1860 to 1920, the rate of gravel extraction may have balanced potential accretion supplied by littoral transport. The 'Point of the Deep' was also stable in depth from 1870 to 1908, but experienced shallowing thereafter (Halcrow, 1999). Progressive beach retreat had converted the site of Hurst Castle into a salient by the early 1960s (Halcrow, 1982). Zig-zag timber breastwork and groynes were installed in front of the Castle in the mid/late 1960s, but were only partially effective as a conservation measure. Furthermore, they promoted downdrift sediment starvation, thus increasing the rate of recession of Mean Low Water (Hydraulics Research, 1982; Halcrow, 1982; Nicholls, 1985).
5.3.6 Morphodynamic behaviour, 1980 to 1996
A sudden increase in the frequency and magnitude of overtopping, overwashing, breaching and breakdown events occurred in the early 1980s, preceded by one major overwash in 1962. Crest lowering, breaching and rollback occurred in 1981/2; 1984/5; 1989/90 and 1994 thereby significantly reducing the resistance of the spit structure to breaching by extreme water levels and wave heights at greater return frequencies (Dobbie and Partners, 1984; Wright and Bradbury, 1994; Wright, in Bray and Hooke, 1998; Mackintosh and Rainbow, 1995; Bradbury, 1998; Bradbury et al., 2009, 2011).
Rollback (landward recession) of 10-25m took place over a length of 2,300m on 29 October 1989. During a few hours, a 1 in 100 year storm on the 16/17 December 1989 caused overwashing to lower the crest of the western sector by 2.5m and displace it landwards between 60-80m. Some 50,000 tonnes of gravel were moved across the crest and back slope onto the marsh and infilled Mount Lake channel via large overwash fans. A further 50,000 tonnes was moved seawards, resulting in a total volume loss 20-30 times the annual average for the previous decade. This recession exposed some 600m² of foreshore to erosion. Emergency reconstruction used 25,000 tonnes of gravel from the site and another 20,000 tonnes imported from external sources. However, erosional scour of the foreshore during the storm; shearing and settlement in the weakly consolidated basement sediments below the new site of the barrier and the scale of displacement meant that the stabilised spit was 12m landwards of its previous position. These events were carefully observed and measured (Bradbury, 1998) demonstrating also the important role of landward seepage at the interface between basement materials and overlying barrier sediments.
The distal sector experienced relatively much less dramatic morphodynamic change over this period. Beach width east and west of Hurst Castle marginally increased between 1984 and 1989 (New Forest District Council, 1990; Halcrow, 1999), and there was also net accretion over some 200m updrift of Hurst Castle. However, after about 1990, it appears that defence measures at Hurst Castle caused local erosion due to outflanking immediately downdrift, with overall advance of Mean High water further north. Net accretion between 1991 and 1995 was 12,000m³ per year for this sector as a whole.
A major storm was experienced on 1 April 1994, causing overwashing, elevation lowering, crest cut-back and fans extending up to 26m across the back barrier slope (Bradbury, 1998). This event emphasised the urgent need for a comprehensive scheme of barrier spit stabilisation. Various mathematical and physical modelling investigations into forcing conditions, and evaluation of the relative merits and probable performance of alternative protection measures, had been initiated in 1990 to this end (New Forest District Council 1990; 1992; 1996; Mackintosh and Rainbow, 1995; Posford Duvivier, 1992; HR Wallingford, 1992 and 1993; Wimpey, 1994; Bradbury, 1998).
5.3.7 The 1996 Stabilisation Scheme; Post-scheme morphodynamics
The main measures that were implemented (New Forest District Council, 1996; Bradbury and Kidd, 1998; Bradbury, 1998; 2009) as the initial phase of the 50 year Hurst Spit Stabilisation Scheme and Beach Management Plan were:
- Recharge, particularly the most vulnerable 800m at the proximal end, with 295,000m3 of 80% gravel and 20% sand dredged from The Shingles Bank immediately offshore that closely matched the median grain size of the prototype sediment (gravel content with a median diameter of 16mm).
- Reprofiling, to include expansion of crest height and width, highest along the proximal sector to ensure that any future overtopping will add deposited gravel at the crest. Dimensions allowed for ground settlement due to compaction and possible shearing affecting the basement sediments. Final form was a crest height of 7mOD at the western end, reducing to 5mOD at Hurst Castle; this was designed to resist predicted overwashing under storm surge conditions (significant wave height of 3.3m) occurring up to a 1 in 100 year return frequency.
- Construction of a short, obliquely-aligned, detached and armoured rock headland breakwater, with a crest height of 2mOD, opposite the position of Cut Bridge. The purpose was to "smooth" the former transport discontinuity (a persistent erosion “hot-spot”) between the terminal end of the rock armoured sector and the adjacent downdrift unprotected beach by dissipating and spreading wave energy. The breakwater creates localised wave diffraction and promotes sand and gravel deposition in its lee, producing a tombolo-like connecting form. Constraints of a steep foreshore slope (deep water) and low strength basement material precluded construction further offshore. Sediment infilling between the breakwater and the beach toe, whilst increasing the downdrift longshore transport rate, would have the potential to be "flushed out" under high wave energy conditions.
- Building a new, large rock groyne west of the Castle frontage, with the main objective of limiting beach volume loss at Castle Point. This measure included the creation of a perched beach and revetment around the Castle walls, a direct replacement for earlier (1980s) defences, and functioning as a headland structure.
- Restoration and strengthening of the rock armour at the far proximal end, involving also a reduction of its gradient.
It was not the intention that, following scheme completion, Hurst Spit would be stabilised as a static structure. Design parameters ensured that it would continue to function as a dynamic barrier system, but providing an enhanced standard of defence (Bradbury, 1998; Bradbury and Kidd, 1998). Periodic recycling of gravel from North Point to the vicinity of Castle Point has been undertaken, at a yearly average of 1,500m³. Crest elevations and widths were designed to be variable in the original design, in response to the spatial variations of wave climate; these are subject to some adjustment as the structure responds to forcing events and settlement (see below).
Since scheme completion up to 2009 Hurst Spit successfully resisted over 20 storms that would otherwise have caused overtopping or overwashing (Bradbury, Cope and Prouty, 2005; Bradbury et al., 2009; Bradbury and Mason, 2009). An exception was both overwashing and crest cutback on 28 October 1996 when wave run-up elevations were higher than predicted. Overwashing alone was experienced in response to two other storm events. Crest cut-back has occurred on several occasions when the design threshold of a significant wave height of 3.5m, with a return period of 1:100 years, was attained or exceeded. This return period has been revised to twice per year (Bradbury, 2008; Bradbury et al., 2009), as the original design wave climate which was based on hindcasting from 15 consecutive years of wind data at Portland has proved to be uncharacteristic of storm events between late 1996 and 2008 (Bradbury et al., 2009; Bradbury and Mason, 2009). Despite wave conditions being more severe than those expected from pre-scheme model predictions, the integrity of Hurst Spit has been maintained. Beach planform resulting from the presence of the offshore breakwater has evolved close to model predictions, largely eliminating the previous transport discontinuity (Bradbury et al., 2009). During early post 1996 storm events crest cliffing also a feature before fine sediment in recharge material was winnowed out or moved down-profile (Bradbury, 1998). Between 1996 and 2000, there was a trend towards steepening of the lower foreshore slope, and net retreat of the upper foreshore (New Forest District Council, 1997-2001) but modified by subsequent recycling and reprofiling. Recycled sediment has been placed on the lee slope rather than the seaward face, the intention being to maintain a dynamic equilibrium profile. Accretion behind the breakwater occurred as predicted, though the rate of accumulation is not known. Overall, the morphodynamic behaviour of the spit when unimodal waves operate has been close to model predictions, with relatively little reduction of crest height due to profile adjustment, "winnowing" of fines, recycling, crest trimming and settlement. During over 40% of storm events up to 2009, when bimodal wave conditions (combined swell and locally propagated wind waves) operated (with significant wave heights in excess of 2.4m) wave periods were some 20% lower than anticipated in the design wave climate, and wave heights, steepness and run-up were consistently higher than expected (Bradbury et al., 2009; Bradbury and Mason, 2009). Beach cross-sectional response to bimodal spectra has on several occasions performed worse than under the unimodal wave climate utilised by the design model, exemplified by overwashing during the storm of 3 November 2005 (Bradbury, et al., 2007, 2009 and 2011; Mason et al., 2008). However, between October 2006 and April 2007 bimodal waves were incident between 5 and 22% of the time (as recorded by the Waverider buoy offshore Milford) without drawdown reducing beach volume-as determined from regular profile monitoring- to below the “alarm” value of 400,000m³. Approximately 40% of storm events, 2003-2008, when significant wave height exceeded 2.4m, were characterised by bimodal wave spectra (Bradbury et al., 2009).
Over the distal sector of the spit, regular topographic surveys at Castle Point have revealed an erosional trend of just over 1,000m³ per year, 1996-2000 (New Forest District Council, 2000; 2010). This has been concentrated along the southern beach face, with a small accretion "fillet" downdrift of the protection works. Since 1997 it has been moving northwards towards North Point (New Forest District Council, 2000; 1997-2001). Since 2001, excess material from North Point has been recycled to Castle Point every three to five years.
Research by Bradbury (1998; 2000) and Bradbury et al., (2005; 2009) into the relations between gravel barrier morphodynamics and hydraulic conditions was fundamentally based on extensive field and model tests carried out on Hurst Spit. This work has demonstrated the critical importance of antecedent barrier geometry in controlling behaviour under extreme forcing conditions; it has also identified an empirical "dimensionless predictive inertia parameter" for defining threshold conditions leading to barrier breakdown. This knowledge is of considerable value in assessing and predicting the future response of Hurst Spit (and other gravel barrier spits) following modification of its natural evolution (refer to Bradbury et al., 2005; Cope, 2005 and Bradbury et al., 2009 for critical discussion of predictive formulation and field measurements of beach topographic responses to nearshore wave conditions and water levels above defined thresholds).
The substrate underlying Hurst spit consists of partially consolidated saltmarsh sediments, silt-filled buried channels and interstratified relict beach gravels with interstitial sand. These record the former seaward basement and recent history of barrier recession. It was considered prior to the 1996 stabilisation scheme that the overburden of recharge material would result in settlement of about 1m over 10 years (with spatial variation) because of their low shear strength. Measurements up to 1999 indicated up to 0.5m of settlement, most of it in the first year (1996 to 1997). It has not been possible to record further movement since, but it is thought to have been negligible (Bradbury et al., 2009).
5.4 The Shingles Bank
This extensive shoal, described by Velegrakis (1994) as the West Solent Ebb Tidal Delta has a main axis of 6.5 km that trends approximately north-east to south west. It has two distinct components, the elongate form of The Shingles Bank proper and, in the north, North Head Shoal. There is considerable spatial variation in morphological form, especially crest height, but both components are asymmetrical in cross-section. The steepest slope of The Shingles Bank is along its eastern flank and is clearly delimited by the Needles Channel; the steeper slope of North Head Shoal is along its north-eastwards facing margin. The Shingles Banks system comprises mobile sand and shingle and is the major offshore sediment sink in Christchurch Bay (Bradbury, 2003).
The dominant sediment type composing both shoals is relatively poorly-sorted gravel. Velegrakis and Collins (1993; 1994) state that vibrocore and surface sampling suggest that flint gravel accounts for a minimum 70%, with various grades and textures of sand contributing 20-25%. Silt and clay, and finely-comminuted biogenic (shell) debris, together constitute less than 2%. HR Wallingford (1994) propose a composition of 90% gravel, 9% sand and 1% silt derived from grab samples taken in the area up to 1km seaward of Hurst Spit. Seismic reflectors reveal that much of the Shingles Bank rests on an eroded bedrock surface (Velegrakis, 1994), and has a variable mean thickness of 2-8m; a maximum thickness of 1m is achieved over a part of the north-eastern sector. This data has provided an estimation of present volume to be 40 to 60 million m³.
A pre-dredging assessment of bathymetric changes in Christchurch Bay, based on historical chart analysis (1882-1988), identified the Shingles Banks system as highly dynamic (Bradbury, 1992). Net growth of more than 3 million m³ was measured for the analysis period, although large-scale spatial and temporal variations in patterns of erosion and accretion occurred, and the accuracy of surveys is uncertain (Bradbury, 2003). Pre-dredging chart analysis demonstrated that the total volume of the system had generally grown since 1882. Subsequent bathymetric surveys (1996-2002) show that rapid changes have continued to occur (Bradbury, 2003).
Analysis of hydrographic charts for the period 1880 to 1968 suggested that the Shingles Bank, as a whole, was accumulating approximately 29,000m³ per year (Lacey, 1985; Nicholls, 1985). Velegrakis (1994) re-examined this data, and analysed subsequent charts up to 1990. From this, he has identified considerable variation in both planform and morphology. In the last decade of the nineteenth century, the Bank had two areas with distinct crestal orientations, but North Head Shoal was barely apparent. This feature was beginning to take on a separate identity by 1921, but at this stage a single crestal orientation prevailed. Over the next 60 years there were non-periodic variations in the extent of separation of North Bank shoal and the Shingles Bank, but there was a distinct trend for (a) the eastern flank to encroach on the Needles Channel; (b) profile asymmetry to be accentuated. By 1979, the Shingles Bank was morphologically unified, with apparent infilling of the channel separating it from North Head Shoal. However, over the next 10 years, this channel became strongly re-established, although the bank complex continued to move eastwards. The analysis of detailed changes from hydrographic charts is limited by both survey accuracy and revision intervals. New Forest District Council (1992) undertook an independent analysis based on digitised hydrographic charts back to 1882 which revealed complex spatial variations of morphology that were interpreted as the product of unsteady processes of both accretion and erosion. This study revealed that the southern area of the Shingles Bank was relatively more stable, with no clear evidence there of sustained eastward movement. The earlier work of Lacey (1985) and Nicholls (1985) was confirmed, with net accretion of 3,226,000m³ between 1882 and 1988 (30,400m³ per year). During this 100 year period, decadal rates of accumulation varied between more than 3% to less than 1% of total volume; between 1972 and 1988, 183,000m³ were added (0.46% of volume calculated for the preceding 25 years). Annual hydrographic surveys of the area dredged for the replenishment of Hurst Spit in 1996 (New Forest District Council, 1997-2001) revealed accretion to be a continuing trend, with almost exactly 1 million m³ added over the 5 years up to 2000. This high rate of accretion is unlikely to be sustained, as it is in response to the 300,000m³ of gravel removed in 1996, and related morphological disturbance). Survey results also revealed complex morphological change, with the persistence of the channel separating North Head Shoal from the main body of the Shingles Bank. Overall westwards migration of North Bank shoal continued up to 2001, with a degree of fragmentation in the central area. The width and depth of North Channel fluctuated annually in precise width and depth, indicating a continuation of spatially variable erosion and accretion. The south-eastern extremity of The Shingles Bank was described by Velegrakis (1994) as building out as a "fan", with some possible transfer of sand to Dolphin Bank.
Despite evidence for fluctuations in planform, and inferred transport paths, all investigations conclude that The Shingles Bank has an overall clockwise movement of sediment (Dyer, 1970; Velegrakis and Collins, 1994; Velegrakis, 1994; Brampton et al., 1998; Halcrow, 1999; Bradbury et al., 2003).
If the historically recent rate of accretion is extrapolated retrospectively, the age of the Shingles Bank cannot exceed 2000 years. However, most researchers are agreed that it started to form no later than mid-Holocene times, following the opening of the western entrance to the Solent and the progressive elongation of ancestral forms of Hurst Spit (Nicholls, 1985; Nicholls and Webber, 1987a; Velegrakis, et al., 1999). A substantial part of its original (and contemporary) volume derived from wave reworking of Pleistocene fluvial gravels. The present-day Pot Bank gravel deposit might originally have been a constituent extension of Shingles Bank, but was detached from it by the incision of the Needles Channel (Webber, 1977). This resulted from erosional scour by strong ebb tidal currents exiting Hurst Narrows. A crude estimate of the net rate of accretion since circa 6,500 years BP (the most recent possible date for the initiation of Hurst Narrows) is 7-9000m³ per year. Apart from supply from the terrace and flood plain deposits of the River Solent, input has come from erosion and recession of the cliffed coastline of Christchurch Bay via the beaches and Hurst Spit. Net onshore transfers and offshore transport out of Hurst Narrows represent potential outputs and inputs. In this sense, the Shingles Bank is a dynamic store; but the fact of long-term increase in volume (net accretion) also justifies its classification as a sink. There remain considerable uncertainties concerning the components of its sediment budget, which can only be clarified by more detailed research on sediment mobility and transport pathways. There would appear to be an input to the southern end of the Shingles Bank via Dolphin Bank (Velegrakis and Collins, 1993; Velegrakis, 1994) but present evidence suggests that it involves sand, rather than gravel. Further north, gravel transport towards the western flank would appear likely.
5.5 Dolphin Bank
Dolphin Bank is made up of predominantly fine to medium, well-sorted, sand. It is some 7km in length, 1.4km broad (at its maximum breadth) and has an elevation of up to 14m above the adjacent, almost featureless, seabed. Its average thickness is 8m (Brampton, et al., 1998). Its eastern extremity is separated by only a narrow corridor from the southern end of Shingles Bank; by contrast, its eastern limit is clearly separated from Dolphin Sand, to the west Seismic data suggests that it rests on an erosional surface cut transversely across Eocene bedrock and coarse (pre-Holocene) superficial sands.
Dolphin Bank has an uncertain depositional history, and may pre-date Holocene sea-level transgression. As a site of net accumulation, it must be classified as a sink. However, the evidence for considerable sediment mobility under present-day hydrodynamic conditions gives it some of the characteristics of a sediment store.
6. Coastal Defence and Habitat Interface Issues
Halcrow (1999) provide a comprehensive review of the conservation and habitat attributes of the Christchurch Bay coastline. This has been supplemented, for Hurst Spit, by the Solent CHaMP (Posford Duvivier and the University of Portsmouth, 2003) and more recently by the Poole and Christchurch Harbours Solent Dynamic Coast Project (Cope and Mortlock, 2012).
Much of this coastline has high national and international importance for geological (stratigraphical and palaeontological) and geomorphological features. Examples include the Eocene rock succession between Hengistbury Head and Becton Bunny, and both Hurst and Mudeford Spits. Nonetheless, considerable loss of intrinsic interest resulted from cliff stabilisation schemes in the 1960s and 1970s. Measures to protect and enhance what remains have been put in place at Hengistbury (Turner, in Bray and Hooke, 1998b; Bray, et al., 1996), whilst innovative "soft" control structures have been proposed for Naish Cliffs (Mackintosh and Rainbow, 1995; HR Wallingford, 1995). The Hurst Spit stabilisation scheme (Bradbury, 1998; Bradbury and Kidd, 1998; Wright, in Bray and Hooke, 1998b) has maintained as much as possible of its morphodynamic character as a barrier structure.
Most of the cliffline is devoid of major biological interest, either because its mobility is inimical to habitat development (e.g. Naish cliffs, and the cliffline east of Barton-on-Sea), or because of protection structures. Exceptions are provided by (i) the vegetation of the cliffs fronting Highcliffe, which have been deliberately planted with an "engineering sward", to enhance their stability (Tyhurst, in Bray and Hooke, 1998b), and (ii) the vegetated cliffline between Sandhills and Friar's Cliff, Mudeford. The latter is largely artificial, dominated by species that have escaped from cliff top properties, notably the Highcliffe Castle estate. There are, however, plants relict from formerly more natural conditions, including those that colonised previous dune habitats near Sandhills. There are opportunities to introduce a more varied cliff vegetation community at Barton, perhaps drawing upon experience gained at Highcliffe. Between Hordle and Milford-on-Sea, relatively stable sectors of cliffline support a patchy vegetation cover of species tolerant of dry, saline conditions. Its local ecological value should be maintained.
Dune vegetation, dominated by Ammophila arenaria (Marram grass) is locally important on Mudeford Spit, where its survival is encouraged by both present and intended management measures (Christchurch Borough Council, 1999). More attention might be given to the introduction of sub-dominant species. This was, up until the early 1940s, a much more extensive habitat, but has been greatly reduced by the loss of the former extent of the spit and by intensification of recreation pressures in an area where there is no direct control of visitor pressure.
Vegetated shingle is relatively impoverished, except along the backshore between Hordle and Rook cliffs and associated with the 'fossil' recurves of the distal part of Hurst Spit. The recession of the back barrier slope of the western and central sectors of Hurst beach has inhibited the establishment of vegetation, but this may have some potential following stabilisation in 1996.
The maintenance of the integrity of both Mudeford and Hurst spits is, of course, crucial to the continuing survival of the rich variety of habitats in Christchurch Harbour and the North-West Solent (Keyhaven and Pennington Marshes) respectively. In the latter case, this is one of several objectives justifying the Hurst Spit stabilisation project, with an expected 50 year life. There may, however, be some loss of both intertidal and terrestrial habitats if this barrier structure continues to evolve by transgressing landwards and its far distal point recurves into Keyhaven creek.
Christchurch Harbour is the site of both high and low saltmarsh (52ha in 2008); wet meadows, dry grassland and both Phragmites and Scirpus reedbeds (8 ha in 2008). Saltmarsh here is unusual for south coast estuaries in that there is a diversity of species, with Spartina anglica failing to gain invasive dominance. The Poole and Christchurch Harbours’ Dynamic Coast Project (Cope and Mortlock, 2012) note that saltmarsh loss since the 1970s in Christchurch Harbour is relatively minor compared with Poole Harbour and other saltmarsh sites across the Solent region. Quantitative analysis was not undertaken given the high degree of geo-rectification error and low resolution of the historical aerial photography.
9km (52%) of the coastline in Christchurch Harbour is protected from flooding and erosion (Royal Haskoning, 2011). Coastal squeeze to saltmarsh over the next 100 years is greater than the potential for habitat re-creation as a result of the Poole and Christchurch Bays Shoreline Management Plan (2011) policies (Cope and Mortlock, 2012). In addition, there are designated landward habitats requiring replacement.
Present policy in the Poole and Christchurch Bays Shoreline Management Plan (2011) is to maintain the balance between different communities and habitats, an approach that is vitally dependent upon shoreline protection against potential breaching at Double Dykes and Mudeford Spit.
7. Knowledge Limitation and Monitoring Requirements
The coastline of Christchurch Bay has been the subject of a number of research projects, and monitoring programmes, over the past 50 years. As a consequence, there is sound qualitative knowledge of most aspects of the coastal process regime, supported by several quantitative studies. Future research and monitoring (some of which is part of commitments by the district authorities and Environment Agency) should emphasise:
- Beach morphodynamics: it will be necessary to continue to quantify on-going volume losses or gains at critical locations, and to determine if profile steepening is a continuing trend. Understanding of cross-shore beach behaviour at locations of recent renourishment, or where there are robust defences, is especially important. Extension of studies to understand beach morphodynamics beyond MLWS is also recommended.
- Longshore sediment transport: there is scope to undertake detailed fieldwork for both groyned and ungroyned short sections of beach. This is best achieved over time periods of not less than one year, to reveal seasonal changes. Ideally, measurements of beach volume should be completed at frequent intervals, so that the effects of different incident waves can be assessed; it might also be possible to observe evidence for counter (i.e. east to west) beach drift, and possibly assess its quantitative significance. Monitoring using tracers could be a technique to consider.
- Net shore-normal onshore and offshore sediment movements in the foreshore and nearshore zones have yet to be established with certainty. At present, they represent inference from analysis of beach volume changes and intermittent hydrographic profiles. It is, perhaps, ironic that there is more uncertainty concerning sediment mobility and circulation in the shallower, inner parts of Christchurch Bay than for offshore transport in the outer area of banks and shoals.
- All of the circumstantial evidence available suggests that the sediment budget of Christchurch Bay is more complex than previously realised. Most components are experiencing either net losses (e.g. beaches) or net gains (e.g. the Shingles Bank). Whether these are in balance has yet to be determined. Inputs via the Western Solent and - arguably - via sources seaward of Dolphin Bank have either not been quantified or adequately measured. An overall loss from the system has been implied from various studies, presumably exiting via a south-westwards pathway. This needs to be resolved by repeat seafloor bedform mapping and analysis of sediment samples.
- Knowledge of sediment mobility and net transport pathways, in the nearshore and offshore zone - particularly in the areas of harbour entrances, sandbars, banks and shoals - mostly derives from ad hoc survey work. Detailed, repeat multibeam and singlebeam bathymetric mapping, maintained by routine re-surveys, would provide a much improved database for future research on wave and tidal current transport, especially in support of numerical modelling and budget studies. Surveys would ideally extend beyond MLWS to capture the interaction of the nearshore area with the beach.
- The difficulties in sediment budget calculation highlighted the need to accurately quantify the input volume to the cell at Long Groyne, output volumes at Hurst Spit, and interaction between the beach and nearshore zone. Identification of direction and volumes of sediment transported from the end of Hurst Spit to where it is finally deposited would also be of substantial interest.
- Information regarding the accurate volume of Shingles Bank, in addition to an understanding on how it changes over time would be of interest. The vast sediment sink has potential as a source of sediment for dredging and consequent use for beach recharge schemes in the local area, especially as the sediment matches the desired properties required. A greater understanding of how potential dredging might impact the bathymetry and sheltering effect of the bank is also of interest.
