About the Study

SCOPAC Committee

Chairperson Councillor Mrs M Penfold MBE, West Dorset District Council.

Vice-Chair Councillor Jackie Branson, Havant Borough Council.

Technical assistance provided to Councillors by Mr Lyall Cairns (Southern Coastal Group Chair) and Dr Samantha Cope (SCOPAC Research Chair).

The STS 2012 update

The 2012 update of the SCOPAC Sediment Transport Study (STS) was funded by the Environment Agency under FDGiA, grant number LDW 41230, with additional contributions from SCOPAC.  

It is referenced as: New Forest District Council (2017). 2012 Update of Carter, D., Bray, M., & Hooke, J., 2004 SCOPAC Sediment Transport Study, www.scopac.org.uk/sts.

Sediment Transport Study 2012

Poole Harbour

1. Introduction

1.1 Origin and Evolution

In common with all tidal basins and large coastal indentations on the south coast the shape and form of Poole Harbour estuary is a result of Holocene sea-level rise that inundated a system of river and stream valleys. This is considered to have reached its climax 5,500 to 6,000 years BP (Before the Present), and May (1969) has published a map to show the probable outline of the shoreline at that time. Several authors (Reid, 1903; Bird and Ranwell, 1964) have argued that Poole Harbour results from the flooding of part of the middle course of the ancestral "Solent River". Velegrakis, et al., (1999) have suggested that the upper part of this river - effectively a 'proto' Frome - was diverted to a south-eastwards course, via a now infilled palaeovalley incised to a depth of at least 12m, in the Devensian cold stage of the Quaternary. It has been suggested that at this stage the estuary was drained via a channel to the north of the present day entrance channel that is across what is now the Sandbanks Peninsula; definitive evidence in favour of this hypothesis is lacking (May, 2005). The tapering headlands, especially on the southern shoreline, and the harbour islands (Photo 1), are remnants of interfluves once tributary to this ancestral river. The cumulative length of the modern, indented, coastline is slightly over 100km. There has been limited interpretation of stratigraphical evidence to elucidate this history of Holocene marine transgression. Godwin, et al., (1958) describe a fresh-water peat horizon at -12.8mOD at Lower Hamworthy that suggests that relative sea and land levels oscillated during the progress of the lower to mid Holocene. Edwards (2001) has provided a more detailed narrative of postglacial sea-level change, revealing from biostratigraphical evidence provided by sediments overlying bedrock in Newton Bay and the Arne Peninsula that overall recovery has been characterised by several successive stages of rising, falling and stable relative regional sea-level. Both Jarvis et al., (1992) and Edwards(2001) concur that sea-level was lower, by about 1.8m, than today in the Iron Age and up to the end of the third century AD. This stage appears to be represented by an erosional step, or microbluff, now buried beneath later mudflat and marsh sediments (refer to May (2005) for a critical assessment of this conclusion). Heath (1988) has demonstrated that Poole Harbour basin is subject to slow subsidence, and this factor must have determined local relative sea-level rise in recent millennia. Long et al., (1999) suggest 0.59mm per year for the late most recent four millennia, whilst Edwards (2001) concludes that the rate has been somewhat less, at 0.53mm per year.  

The original (post mid-Holocene) planform of the harbour has been substantially modified by both erosional and depositional processes. Erosional backwearing of low cliffs, exceptionally approaching 10m in height, has taken place at a number of localities exposed to the modest but not insignificant capability for wave propagation within the harbour. Accretion has involved the narrowing of the harbour mouth as a result of spit growth generating both the Sandbanks and South Haven Peninsulas. Originally over 3.5km in width, it is now reduced to approximately 350m (Photo 2). Several of the now degraded bluffs, most of which are less than 5m in height, around the harbour margins probably owe their origins to earlier conditions of greater exposure to the more energetic wave climate of Poole Bay (Photo 3). It is also likely that waves would have driven some sands from Poole Bay into the harbour over this period. A much more important contribution from accretion has been the creation of extensive mudflats, particularly as a result of their colonisation and stabilisation by salt marsh vegetation. Saltmarsh growth was especially important in the first 30 years of the twentieth century as a result of the appearance and rapid invasion of Spartina (Photo 4), although since the 1930s there has been a progressive reduction in the area of saltmarsh (Gray and Benham, 1990; Gray, et al., 1991; Raybould, 1997; Born, 2005; and Cope and Mortlock, 2012). The documentary evidence recording both the growth of the South Haven Peninsula since the late 16th Century (Diver 1933) and the details of Spartina anglica invasion and retreat indicates that physiographical changes within the historical period have been very considerable (A and C Archaeology, 1991). The narrow confinement of the mouth of Poole Harbour relates to the extension of the Sandbanks and South Haven spits, probably involving a complex history of barrier progradation in relation to sediment supply from Poole Bay and sea-level oscillations over at least the most recent three millennia. This critical stage in the estuary’s evolution awaits clarification (Cook, 2007).

The contemporary area of Poole Harbour, at maximum high water, is close to 3,600ha. May (2005) calculates that this is approximately 1,000 ha less than the area at about 6,000 years BP, when the most rapid stage of postglacial sea-level rise concluded. Whilst much of the shoreline of the southern part of the estuary and islands has not been modified by defence structures and other development, the northern, north-eastern and north-western harbour boundary is mostly fixed in position by seawalls, wharves and banks. One third of the defences around the harbour perimeter are privately owned and maintained. The almost landlocked embayments of Lytchett and Holes Bays on the north shoreline are occupied by mudflats and marshes that are emergent at low water, but their margins are partly developed and protected.

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.channelcoastal.org. In 2012, an extensive high resolution, 100% coverage swath bathymetry dataset was collected in Poole Bay, through the Maritime and Coastguard Agency’s Civil Hydrography Programme (this survey did not include the area at the harbour entrance and approach channel within Harbour Authority limits). The Southeast Regional Coastal Monitoring Programme commenced in 2002. The Lead Authority is New Forest District Council, with data collection, analysis and reporting led by specialist teams at the Channel Coastal Observatory (CCO), Canterbury City Council and Adur and Worthing Councils. (See CCO Annual Survey Reports for further details).

1.2 Tidal and Wave Conditions

The tidal regime of Poole Harbour has been described by various authors (e.g. Bird and Ranwell 1964; Poole Harbour Commissioners 1982, 1985; Gray 1985; Hydraulics Research 1990, 1992; HR Wallingford 1995, 1996; Halcrow, 1999; May, 2005; Royal Haskoning, 2011). It is characterised by a small double high water effect that is transmitted from Poole Bay, with a mean tidal range of 1.8m at springs and 0.6m at neaps (actual values vary with location). Tidal levels are above mean from about 2 hours after Low Water to approximately 2 hours before the next Low Water, i.e. for nearly 16 hours per day. Predictions of the timing of High and Low Water are not infrequently modified by prevailing meteorological conditions. All tidal oscillations are recorded, on average, about 45 minutes later at the mouth of the Frome compared with the entrance to the harbour,  but viewing the estuary as a whole tidal variation is small, and at standing high water Poole Harbour is analogous to a lagoon. Ebb tidal stream velocities are higher than those of the flood in the lower estuary, with maximum speeds of approximately 2m per second at the harbour entrance. In contrast, the upper estuary is dominated by relatively weak flood tide velocities. Tidal flows close to the harbour periphery are also weak, except at the constricted entrance to Lychett Bay, and at the Town Quay. Characteristic velocities in the Main Channel are 2m per second with lower flows of 1 to 0.1m per second in the shallower central valley (Royal Haskoning, 2011). At Low Water, the tide withdraws to the main channels, when the total water volume is about 23 million m³ (Poole Harbour Commissioners 1985). The tidal prism has been reduced over the past 200 years as a result of marginal land claim and marsh and mudflat accretion; this has caused the mean water level to rise, in the order of 20 to 25cm, and the main channels to deepen (May, 2005).  There is a small input of water volume from the four main rivers draining into the harbour, calculated at a mean of 475,000m³ for a single tidal cycle (2% of the total volume). Much of this may be confined by the weak saline wedge that differentiates the water mass in the upper harbour (Green 1938; Halcrow, 1999). Extreme water levels (mOD) are given in Halcrow (1999), and reveal only small differences between various locations. Contemporary relative sea-level rise (1955-1997) has been calculated at 4mm per year but there are some doubts about the reliability of the tide gauge data for Poole Town Quay (Pethick, 1993; Halcrow, 1999) Hubbard and Stebbing (1968) calculated that sea-level rise at Keyworth was 1.85mm per year between 1912 and 1966, but the same reservation must apply.

Halcrow (1999) have modelled extreme wave heights based on hindcasting from local and regional wind data, applied to a detailed bathymetric grid. These vary from 0.5 to 1.2m for a 1 in 100 year recurrence, depending on location with respect to wave fetch. Northern and eastern parts of the harbour tend to be the most energetic for they are exposed to long fetches that coincide with dominant west and southwesterly approaching waves. The model showed that storm waves at the entrance do not penetrate into the main harbour, mostly due to diffraction and refraction effects at, and seawards of the entrance channel. The wave climate is thus dominated by depth-limited, locally-generated waves, whose average heights vary between 0.2m (south and west shores) and 0.2 to 0.5m (north and northeast shores).

2. Sediment Inputs

2.1 Marine Input

Input from Poole Bay

Research carried out on sediments, and on sediment transport pathways, within the harbour has been unable to identify a specific input from Poole Bay due to similarities in sediments eroded from local cliffs with those available within Poole Bay (Colenutt, 1994). Investigations on the mobility of sand in the harbour entrance and the Swash Channel immediately seaward (Hydraulics Research 1986, 1988, 1991; HR Wallingford 1993b; HR Wallingford, 2003) do however indicate an unquantified potential for inward transport during combinations of storm wave and flood tide conditions. Indeed, the presence of sand and gravel flood tidal deltas immediately inside the entrance (Photo 2) together with the prevalence of fine to coarse sandy sediments covering the harbour bed in the vicinity of the entrance would tend to verify the occurrence of this process. It is thought that there is a net input of suspended materials into the harbour from Poole Bay. This is a function of the ebb dominant tidal regime where the flood tide enters more slowly, but over a longer duration giving greater opportunity for input of fine suspended sediments. The existence of a net marine suspended sediment input has yet to be confirmed by actual measurements.

2.2 Fluvial Input

FL1 River Frome and other Fluvial Input

River-transported sediment is delivered to the harbour by two major rivers (Frome and Piddle) two smaller rivers (Sherford and Corfe) and a number of small streams that drain tributary catchments underlain by erodible sands and clays. The lower Frome, in particular, is characterised by a wide alluvial flood plain that represents a sediment store available for re-working; over the long-term, however, fluvial sediment that might have been discharged into Poole Harbour has been trapped as alluvium within this zone. The practice of water meadow management may have enhanced this effect over recent centuries (May 1969).

Hubbard and Stebbings (1968) quote a figure of 36,700kg per day (approximately 720 tons per meter) of suspended sediment load for the Frome at Wareham Quay. Multiplied to give an annual input, the amount discharged over the past 100 years is only about 1/25th to 1/30th of the total sediment accretion that has taken place in the harbour as a whole over this period, as determined from documentary evidence of Spartina-assisted mudflat expansion (Raybould, 1997). The above figure, which is for January 1967, may not be representative, and both Bird and Ranwell (1964) and May (1969) state that maximum sediment yields take place at times of higher than average runoff. Without quoting a source for his data (but believed to be from a gauging station at Wareham), May (1969) states that near bankfull discharge on the Frome operates for 1 day in 20. Bird and Ranwell (1964) described in general terms the situation in Poole Harbour during exceptionally high-river discharge in November 1962, remarking that the, "upper half" of the harbour was "cloudy" with suspended sediment, whilst the "lower half" was clear. Such statements are, at best, ambiguous as sediment samples of mud obtained near the lower end of the Middle Channel revealed no minerals of alluvial origin (McMullen, 1985). Gray (1985) has suggested that suspended sediments introduced by rivers settle out onto mudflats and channel margins in the upper harbour because of the poor flushing characteristics of the tidal regime, which causes the plume of river water to be confined by the near-stationary front of brackish water during standing water. Colenutt (1994) indicated through Particle Size Analysis and X-Ray Diffraction that the spatial distribution and composition of surficial sediments is more closely related to geological history than to present hydrodynamics, with majority of fine-grained sediments within the harbour deriving from in situ resuspension and very insignificant transport from fluvial inputs estimated. Rendel Geotechnics and the University of Portsmouth (1996) estimated that the major rivers discharging into Poole Harbour contribute 1492 tonnes per meter of suspended load and less than 600 tonnes per meter of bedload,  although these values may be over-estimates if they do not exclude fine textured organic debris.

Quantitative understanding of the contribution of fluvial material to the sediment budget of the harbour is poor, although Hydraulics Research (1990) report clay minerals obtained from the north-east sector which could be of fluvial origin. Intuitively, it may be argued that much of the suspended load received at present is temporarily stored by mudflats and at channel margins in the vicinity of the Upper Wareham Channel.

May (1969) pointed out that there have been significant land use changes in the catchments of the rivers discharging into Poole Harbour since at least the early 1930s. On balance, they may have resulted in a slight decrease in the volume of sediment load, although changes affecting the agricultural economy since the late 1970s may have partially reversed this effect.

2.3 Coast Erosion

» E1 · E2 · E3 · E4a · E4b · E5 · E6 · E7 · E8

May (1969, 1976; 2005) has given detailed accounts of processes and rates of cliff development, particularly at Shipstal Point and vicinity. Gray (1985) also provides a well-illustrated, descriptive inventory of all sites with both active and degraded cliffs around the periphery of the harbour. The remaining literature consists of brief inferences or implications of past or present inputs from cliff erosion. The available evidence is arranged here by taking a clockwise circuit of the shoreline, starting at the mouth of the River Frome. Cliffs and shoreline platforms developed in both the Bagshot sandstone outcrop and at the seaward edge of marshes are included, despite obvious contrasts in their scales of development. The saltmarsh surface between +1.3m and +0.05mOD in the southern harbour is truncated by an eroding bluff 0.2 to 0.3m in height (May, 2005). The rapid expansion of Spartina anglica from the late nineteenth century to the mid or late 1920s brought about significant reductions of wave energy at previously cliffed sites. The subsequent "die back" of Spartina has at least partially reversed this trend at several sites (Gray, 1985; Raybould, 1997; May, 1998).

E1 Keysworth Marsh (see introduction to coast erosion)

Both Hubbard and Stebbing (1968) and Gray (1985) report low cliffs (micro-scarps or bluffs) between 4 and 75cm in height at the edge of the marsh; Ranwell (1964) remarked that this and several other marsh margins fringing the Upper Wareham Channel are eroding. May (1969) has confirmed this, and identifies wave action as the cause. Negligible marsh edge recession is evident from analysis of Coastal Monitoring Programme 2008 and 2013 lidar and aerial photography.

E2 Rockley Cliff, West of Lake (see introduction to coast erosion)

Reported and illustrated by Gray (1985), May (1969) calculated an erosion rate of c.0.3mm per year for an unspecified length of the Hamworthy shoreline. The cliff is some 10-15m in height and 1km in length, and comprises sands and some gravel. Analysis of Coastal Monitoring Programme 2008 and 2013 lidar and aerial photography, and baseline topographic data indicates that the width and levels of the local beach have remained stable and minimal cliff recession is evident.

E3 Lilliput to Sandbanks (see introduction to coast erosion)

Eroding cliffs of up to 18m height within Tertiary sands formerly existed at Evening Hill prior to slope stabilisation in the 1970s and renovations in the 1990s. The 2004 arrow has been removed as there is no longer a source from these cliffs due to stabilisation works.

Further to the southeast, Gray (1985) notes that a low bluff cut into saltmarsh existed in front of the sandy-muddy beach along Shore Road (B3369). Erosion proceeded rapidly in the 1980s and 1990s so that very little of the original saltmarsh now exists at this location. In the period 2007 to 2012 a sandy beach developed in this area, which may indicate potential drift convergence. The Borough of Poole Council started a small trial in 2013 attempting to regenerate saltmarsh colonisation at this site, which is monitored through the Coastal Monitoring Programme.

E4a Brownsea Island (see introduction to coast erosion)

Gray (1985) provides illustrative evidence of basal undercutting and gulleying of the cliffed slope along most of the south coast, endorsing May's (1969) calculation that the southwest Brownsea coastline retreated at a rate of up to 0.45mm per year, 1886-1952. More specifically, May (2005) states that the retreat rate for the south coast of the island was 0.1m per year prior to Spartina invasion and the accretion of a protective saltmarsh; since its dieback starting in the 1930s the retreat rate has increased to 0.85m per year. Bamber and Ranger (1990) summarise a consultants' report to the National Trust (1988) indicating severe erosion in the vicinity of the seawall around the northeast quadrant of the island, first built in 1854 and repaired in 1918 and 1927 as well as the defended southeast sector of shoreline. Artefacts of Iron Age 'B' and Romano-British ages have been discovered, apparently in situ, 25 to 30m offshore of the quay (National Trust Guide, 1984; Jarvis, et al., 1992). Erosion is more limited on the north coast and the majority of the steep coastal slopes are vegetated. Active cliff erosion at Portland Hill and Barne's Bottom on the south coast and on the west-facing coast is described by Appleton (1995), which is ascribed as much to runoff and hydrogeological conditions as to basal wave abrasion. Former cliff stabilisation at the western end of the island is thought to have reduced beach volumes along the south coast, thus accelerating coastline recession since the mid-1980s. The dilapidated ad-hoc defences along the west and south coast of the island, consisting of steel sheet piles, Scots Pine piles, gabion baskets, shards and plastic netting were removed as part of the Brownsea Shoreline Restoration Project (2011) in 2011. Analysis of Coastal Monitoring Programme aerial photography (2008-13) shows 2-3 metres of erosion of the cliffs on the west side of the island, approximately 2 metres of erosion of the cliffs to the west of the Scout camping grounds, and small scale landslides on the southeast and west sides of the island. Elsewhere, the island’s cliffs and coastline are comparatively stable.

E4b Furzey and Green Islands (see introduction to coast erosion)

Active block detachment by weathering and shallow mass movement are recorded at two sites on Furzey Island, although no recession rates are given (Pfaff, 1994; Pethick, 1993; Pethick and Forster, 1994). Given similar rock lithology and wave climate to the south coast of Brownsea Island, the causes and retreat rates of these small scale cliffs on Furzey Island are probably comparable. May (2005) estimated a contemporary retreat rate of 0.06m per year for the cliffs on the east coast of Green Island.

E5 Brand’s Bay (see introduction to coast erosion)

Gray (1985) describes both degraded and active cliffs, up to 4m in height, cut into a variety of materials including marsh sediments, dunes and in situ Bagshot Sands (Photo 3). Negligible recession of the low sandy cliffs is evident from analysis of Coastal Monitoring Programme 2008 and 2013 lidar and aerial photography data.

E6 Goathorn Peninsula (see introduction to coast erosion)

Gray (1985) briefly describes and illustrates a sandy cliff up to 10m in height on the western shoreline, observed to be actively retreating due to basal undercutting in the early twenty-first century (May, 2005). Negligible recession of the low sandy cliffs is evident from analysis of Coastal Monitoring Programme 2008 and 2013 lidar and aerial photography data.

E7 Shipstal Point and Vicinity (see introduction to coast erosion)

May (1976) provides a detailed account of spatial variations in cliff form, cut into interbedded Bagshot Sands and "Ball" Clays, and overlying superficial gravels. Cliff morphology primarily reflects control exerted by exposure to high run-up breaking waves, in turn a function of fetch distance and local refraction. Periodic wave action removes basal debris produced by wash, gulleying, small-scale falls and recreational trampling forming local beaches and contributing towards a northward trending spit within Arne Bay (Photo 4). Cliffs to the south of Shipstal Point exhibit no contemporary toe erosion, having been removed from the direct influence of waves by saltmarsh growth in Middlebere Bay. A short-term monitoring programme revealed that sub-aerial processes contributed 25m³ of debris during the period April to October 1971; 20% of this total was produced by storm runoff over the cliff face in two weeks in late September to early October. 70% of this amount was removed by waves close to calculated maximum heights in January 1974. In any context, this is a particularly interesting study of the magnitude and frequency of processes maintaining and modifying cliff form. To what extent it is representative of other sectors of the eroding coastline of Poole Harbour is debatable. This is because of major differences in exposure to wave action due to constant changes in coastline plan and orientation, and spatial variation in the extent of fringing mudflat/marsh development. Minimal recession of the sandy cliffs is evident from analysis of Coastal Monitoring Programme 2008 and 2013 lidar and aerial photography data.

E8 Arne Peninsular (see introduction to coast erosion)

Cliffs are well-developed between Hyde's Quay and Russell Quay, on the north-west shore cut into Bagshot sands and clays. Toe erosion and cliff top recession, small slumps and gulleying accelerated by recreational pressure were mapped by student groups from 1978 to 1981 (Leeson House Field Studies Centre, unpublished). Gray (1985) describes this coastline as "strongly eroding" and May (1969) suggests a rate of retreat of between 0.35 and 0.4mm per year, 1886 to 1952. Gray also describes a strongly defined erosional micro-scarp at the seaward edge of the mudflats and marsh in Arne Bay. Sands and some gravels are contributed to the shore from the cliffs. Negligible recession of the sandy cliffs is evident from analysis of Coastal Monitoring Programme 2008 and 2013 lidar and aerial photography data.

3. Sediment Transport within the Harbour

3.1 Littoral Transport (Shoreline Drift)

» LT1 · LT2 · LT3 · LT4 · LT5

Although the surface area of Poole Harbour, at high water, is large, its irregular and indented planform, together with the presence of one large and several small islands, restricts fetch distances available for wave generation. Waves of modest height break along most sectors of the shoreline and are affected by localised refraction within the shallow harbour waters (HR Wallingford, 2003). This is reflected in the process of relatively weak littoral drift of beach sediments.

The detailed literature is restricted, but all general surveys either report or assume a net eastward component to littoral transport at most locations. May (2005) observes that prior to Spartina colonisation several of the harbour salients on the southern shoreline, and islands, had north-eastwards recurving sand and gravel spits attached to their eastern limits.

LT1 Arne to Patchins Point (see introduction to littoral transport)

Sands and gravels released by cliff erosion (E8) drift northeastward around the Arne Peninsula. This is confirmed by Bird and Ranwell (1964), who describe shingle spit structures along the northeast of the Arne Peninsula that appear to have extended eastwards across the northern part of Arne Bay to form Patchins Point. Saline lagoons and saltmarshes have developed in the shelter created behind the thin spit. Analysis of Coastal Monitoring Programme 2007 and 2011 lidar and aerial photography data indicates negligible change in widths or levels of the narrow fringing sand foreshore.  

LT2 Hamworthy (see introduction to littoral transport)

Bird and Ranwell (1964) identify the shingle beach at Hamworthy as having been derived from erosion of gravel deposits at Rockley cliff to the west. Analysis of Coastal Monitoring Programme 2007 and 2011 lidar and aerial photography data indicates negligible change in widths or levels of the narrow fringing sand foreshore, but sediment accumulations within groyne embayments at Hamworthy Park support a weak less than 1,000m³ per meter eastward drift of sand and gravel. A beach replenishment was undertaken at Hamworthy Park in the late 1980’s early 1990’s using a land based source (Dave Robson, Borough of Poole Council, pers. comm., 2016).

LT3 Eastward drift along Brownsea Island (north and south shores) (see introduction to littoral transport)

Drift appears to operate from west to east along the north and south coasts of Brownsea Island with two distinct pathways originating around Pottery Pier at the western extremity. Gray (1985) and May (1998) note the presence of small forelands and a sandy spit projecting eastwards along the northwest shoreline of Brownsea Island; whilst Bird and Ranwell (1964) and Appleton (1995) observe that pieces of broken pipeware from a nineteenth century pottery site at the southwest of the island's coastline are now incorporated into beaches along the southern and eastern shores. Patterns of spit projection along the eastern coastline of Brownsea, whilst somewhat modified by development, appear to indicate drift junctions.

Analysis of Coastal Monitoring Programme 2007 and 2011 lidar and aerial photography data indicates negligible change in widths or levels of the narrow fringing sand foreshore on both the northern and southern shores of Brownsea Island.

LT4 Shipstal Point: Northern and Southern Spits (see introduction to littoral transport)

The most detailed study of littoral drift anywhere in Poole Harbour is that of May (1976), who has carefully examined the association of cliffs, beaches and spits in the vicinity of Shipstal Point. (See also May, 2005). He observes that the southern spit is migrating southwards, whilst the net direction of longshore transfer of sediment along the northern spit (Photo 4) is south to north. Using serial map and air photo analysis, May concludes that the northern spit has narrowed since the early 1950s, possibly due to the construction of a set of groynes near the proximal end of the feature. However, the distal end shows some evidence for recent broadening and possesses a set of recurves; these characteristics point to the continuity and relative rapidity of longshore drift at this location. The southern spit has appeared only since the 1920s and owes its origin to the colonisation of adjacent Middlebere Bay by Spartina anglica, since 1914. Prior to this, waves impinging along this sector of coastline would have approached from the south, but the growth of Spartina marsh has had a protective effect and has set up wave refraction patterns that cause a net north to south movement of material. The southern spit is in an equilibrium condition, adjusted to prevailing local wave climate, whilst the northern spit is currently a transient form adapting to alterations in the longshore transport sediment budget. The latter may be a function of absolute changes in wind and wave frequencies in recent decades (Halcrow, 1999), as well as the alterations in coastal configuration already mentioned. Overall, May's study provides an excellent illustration of the fact that net littoral drift pathways can operate in opposed directions on adjacent parts of the coastline due to abrupt alterations of exposure and wave fetch that have resulted from the rapid growth and dieback of Spartina marshes. It is also a convincing example of the value of detailed study of historical sources in elucidating the causes and consequences of change.

Analysis of Coastal Monitoring Programme 2008 to 2013 lidar and aerial photography data supports a drift divergence with both the northern and southern spits extending approximately 10m during this period, although changes in widths or levels of the narrow fringing sand foreshore are negligible.

LT5 Whitley Lake (see introduction to littoral transport)

Analysis of Coastal Monitoring Programme 2007 to 2012 topographic data supports a weak northeastward drift of less than 1,000m³m-¹ of sand, as indicated by the establishment of a sandy beach ridge in front of the saltmarsh, although changes in widths or levels of the narrow fringing foreshore are negligible.

3.2 Tidally-induced Transport

There are a number of papers and reports that either directly or indirectly identify transport of sand, sandy-mud or mud sediments at or close to the bed of Poole Harbour without specifically identifying the mechanism(s) of transportation. For the purposes of this section, only implied tidally-induced disturbance and movement will be considered.

Green (1940) observed that bottom samples of fine to medium sand are well sorted, or graded, and are loosely consolidated; the only exception appears to be at and close to the harbour mouth and in areas of relatively rapid changes in sea bed relief. The latter occur in the vicinity of the main channels, i.e. at the boundaries of the main banks or shoals. He interpreted this as implying that good sorting reflected equilibrium adjustment to stresses applied by tidal forces, with mixed sediments indicating greater mobility. He notes that the ebb current is stronger than the flood in the lower, south-eastern parts of the harbour, with the reverse the case in Upper Wareham Channel, thus speculating that there was a "negative sediment balance" in both areas. Green also mentions that the northward-moving flood and eastward moving ebb tidal streams are capable of transporting fine calibre sediment across the Middle Ground and into the Main Channel. Sampling here, and elsewhere, proved the presence of a layer of silt overlying sand, which Green (1940) considered to be the result of mudflat erosion in the upper harbour. The implication is that silt and clay particles had settled out of suspension over several preceding years, and that this material needed to be removed before tidal forces could entrain sand. Green, et al., (1952), reporting the results of a partial resurvey of Poole Harbour in the previous year, provide a small-scale map of an "embayment" of coarser bottom sediment extending across Middle Ground from north of Brownsea Island. This is taken to correspond with the "stronger and most consistent flow of the flood tide". An area of mixed sand size-fractions, extending north-east from Brownsea Island to beyond the axial line of the Main Channel, is considered to result from the reduction of velocity of both the flood and ebb tidal streams. Scour and abrasion of the surface of the former Soldier Ground (an eastward extension of Middle Mud) was implied from the natural widening and deepening of the New Cut, a point also noted in Green (1940). In both papers, the disappearance of Soldier Ground was predicted, and duly occurred by the late 1950s. The narrowing of the Main Channel between 1938 and 1951 might have been the result of the movement of sediment away from Soldier Ground during this period.

McMullen (1985) draws upon Green (1940) and Green et al., (1952), but makes some additional points of interest. Discussing tidal streams, he notes that whilst the spring ebb is stronger in the Main and Wych channels the flood stream has greater velocity compared to the ebb in the Middle Channel. This raises the possibility that net transport is in different directions in different channels. Estimates are given of sediment transport rates, using the Van Rijn relationship applied to mean current velocities. Current velocities subdivided for grain size classes for both the ebb and flood streams in the Main, Middle and Wych Channels range from 1.09 to 0.12m³m-¹. A detailed analysis is not attempted and would not be warranted on the basis of such a limited survey; however, spatial variations in rates were found to be consistent with the pattern of change in bank and channel configuration. Highest values were reported from Wych Channel, the lowest from one location in the Main Channel. The report also carries brief details of sediment samples, the overall pattern of which suggests that, in the harbour as a whole, the sand fraction has become marginally finer since the late 1930s. This may be an artefact of the different sampling techniques used in this investigation and by Green (1940). The observation that Middle Mud is stabilising may be because the pre-existing silt veneer is being, or has been, "washed off", with more resistant sand re-exposed.

Hydraulics Research (1990) conducted field, numerical modelling and other experimental approaches to investigate potential sediment transport in and close to the Middle Channel. This work was commissioned by Poole Harbour Commissioners as a contribution to proposals to realign the main port navigation channel. A map of potential erosion and deposition of sand is given, based on: (i) analysis of detailed changes in the depth and curvature of the channel in recent decades; (ii) theoretical considerations based on a 'best fit power law' equation developed for estuarine transport simulation in the Great Ouse. The point is made that maximum accretion will occur where net transport vectors converge, and an attempt is made to map the contributing sediment pathways. The conclusion that the Middle Channel has a positive sediment balance concurs with earlier studies, but is here demonstrated quantitatively. Some figures of gross "siltation" rates are given, which range from 19,000 to 7,000m³m-¹ for different sample locations. Up to 18,000m³m-¹ may accumulate in the mid-section of the channel, but with scour on the main bend reducing this to 7,000m³m-¹. These values appear to relate exclusively to sandy-silty sediments.

Diver observations of the estuary bed in both deep channels and on sandbank surfaces are given as an Appendix in McMullen (1985). Well-developed rippled patterns in fine sand and mixed sand and silt are reported, but without details of size and spacing. This may be taken as reasonable evidence of the role of bedload transport by tidal currents. However, the quantity and quality of data on pathways and rates of transport does not allow more than a very generalised picture. Some additional detail on the distribution and sorting of sediments within and marginal to the main channels are provided in HR Wallingford (2003), but do not contradict the pattern described above. Taking into account evidence of input, circulation and output near the harbour mouth, it may be tentatively stated that the upper and lower sectors of the harbour have a negative sediment budget whilst the central area is in positive balance. There is, as yet, insufficient evidence to support the contention by McMullen (1985) that sediment transport in the harbour as whole functions as a balance between inputs from fluvial sources with output at its mouth.

3.3 Wave-induced Transport

In Section 5.1 it is implied that sediment from silty-clay through to coarse clastic grades may be entrained by breaking waves along parts of the harbour margins, especially the northern and north-eastern shorelines where the upper beach is composed of either fine gravel or coarse shell debris (Halcrow, 2004). Howard and Moore (1988) observe that fines may be selectively transported in suspension. Laboratory and flume tests on muds sampled from the Middle Channel indicated shear strength values sufficient to resist current-induced erosional forces, but this work included theoretical assumptions on wave orbital velocities and stresses (Hydraulics Research, 1990). Disturbance of fine-grained sediments in a shallow area east of Little Channel (less than 1.4m depth) by waves may be sufficient to prevent deposition of material there (Hydraulics Research, 1992). Boat wash could be a factor everywhere along all principal and tributary channel margins although especially in the vicinity of the Main Channel that is navigated by ferries and larger commercial vessels.

In order to assess the role of waves in both suspending and transporting sediments within the inter-tidal area of the harbour, full-scale wave climate modelling was needed. This has been undertaken for the area of the Middle Channel (Hydraulics Research, 1990; HR Wallingford 1993b, 2003), and provides a basis for further research. Modelling by HR Wallingford (2003) suggests that storm waves can penetrate a short distance through the main entrance channel and contribute coarse sand that is subsequently deposited at or close to co-adjacent sub-tidal seabed and the inter-tidal zones. Nonetheless, the limitations of these models should be noted, not least the lack of adequate wind data used for estimating significant wave-height (Halcrow, 1999; 2004).

4. Sediment Outputs

4.1 Estuarine Outputs

EO1 Wareham Channel

The tidal regime results in the ebb tidal stream having a higher velocity than the flood towards, or just before, low water. In most, but not all, of the main channels including the Wareham Channel, an output of coarser sediment (fine sands and larger) down the main axis might be expected on theoretical grounds (Green 1940, McMullen, 1985, Howard and Moore 1988). The actual transport occurring is, however, likely to depend upon the availability of transportable sediment and it is likely to be limited to the relatively small quantities of coarser sediments delivered by the Frome and Piddle Rivers. By contrast, fine suspended sediments appear to move preferentially up the channel to be deposited on mudflats and at the heads of creeks.

EO2 Main and Middle Channels and Harbour Entrance Channel

The ebb tidal stream has a higher velocity than the flood at the harbour mouth and in most of the main channels. Within the harbour, peak ebb flow generally occurs at, or just before, low water. In most, but not all, of the main channels, an output of sediment down the main axis of the basin might be expected on theoretical grounds. Ultimately, this would translate to a loss of bedload sediments at the entrance that would involve fine, medium and possibly coarse sand transported as bed load (Green 1938, McMullen, 1985, Howard and Moore 1988). It does however assume that transportable sediment is available and that zones of storage do not intercept the transport pathways. The former is a consideration within the Wareham Channel, whereas the latter is important to the Middle Channel pathway. Studies have indicated that sediments move from the Main Channel to accumulate within the Middle Channel and navigation channels (Dave Robson, Borough of Poole Council, pers. comm., 2016), where there is a possible reversal of transport (the Middle Channel is flood dominant, whereas the Main Channel is ebb dominant). The net effect could be for interception of outward moving sediments by a tidal circulation of sediments around sites of net accumulation to the north of Brownsea Island.

Some authors, e.g. Gray (1985) have contended that tidally transported sediment loss is prevented or inhibited by the constriction of the harbour mouth. However, this effect would mostly serve to magnify the dominant ebb current and at the same time offer some resistance to open sea waves pushing sediment into the harbour. The possibility of sediment input by flood tidal streams in combination with wave action may produce a "dynamic sediment balance" (McMullen, 1985), although the net outcome of the various inputs and outputs has yet to be determined.

A detailed survey of Chapman's Peak by Hydraulics Research (1991) has been undertaken, using a 20m gridded tidal vector model, bathymetric survey and sediment sampling. This has suggested a gross siltation rate of 5,000m³m-¹ (accurate to within a factor of 2) where tidal stream vectors converge on the bank. The potential for the variable erosion, transport and deposition of gravel, shells and/or sand by tidal streams is therefore considerable. Whilst this study, located off Sandbanks within the harbour mouth, demonstrates tidally-induced sediment mobility, it does not reveal the long-term sediment balance at the entrance, nor were the effects of wave action considered.  

4.2 Land Claim

May (1969) calculated that since about 6000 years BP, at the conclusion of the phase of most rapid sea-level rise during the Holocene transgression, the area of Poole Harbour has been reduced at a rate of between 0.11 and 0.13 ha annually due to natural sedimentation and a long history of land claim. Using a wide range of documentary sources, which allowed corroboration and independent checking, he gave a figure of approximately 1.6 ha per annum of land claim between 1807 and 1966, including 116 ha from 1920-1966. Thus, something between 20-25% of the original area of the harbour, as measured at High Water Springs, has been converted to non-tidal land (approximately 1000 ha). These figures do not distinguish between natural and artificial causes, though the sources listed in May (1969) are sufficiently accurate for this calculation. The major recent cause of natural land claim has, of course, been the expansion of saltmarshes as a result of the spread of Spartina anglica between 1880 and 1930, though the longer-term contribution of other species should not be underestimated. Hubbard and Stebbings (1968) have used documentary materials to indicate that the siltation and progressive reduction of the area of Holton Mere throughout the nineteenth century preceded the arrival of Spartina anglica (in this case in the early 1920s). Other sites have not been researched at the same level of detail, but they might indicate similar trends. The substantial losses of inter-tidal wetland as a result of Spartina anglica degradation and "die-back", starting in the mid-1920s, has almost wholly offset earlier gains. Gray (1985) calculated that for Holes Bay the total intertidal area has reduced from about 330 ha to 250 ha since 1924 mostly due to urban encroachment. At this site Spartina anglica marshes reduced from 200 ha to 70 ha although some of the areas became replaced by lower level mudflats. In addition, there has been a rapid increase in reedbed extent between 1993 and 2008 (Cope and Mortlock, 2008).

The contribution of artificial land claim has not been quantified for all parts of the harbour margins, though May (1969) gives details for the northern shoreline east of Rockley. The greatest reduction in area due to urban development, port extension, marina construction, etc., has been between Parkstone and Lychett Bays.

4.3 Navigational Dredging

Access to the port of Poole by shipping requires regular maintenance of navigation channels, particularly in the Middle (or 'Middle Ship') Channels. The 350m wide entrance channel is dredged to -10mCD to accommodate shipping routes in and out of the harbour. Changes in the relative positions of banks and shoals have caused some displacement of channels, although there has been general stability since the first reliable and detailed hydrographic survey in 1785. For operational reasons mostly related to increasing vessel sizes and draughts, capital dredging of the principal access channels has been undertaken from time to time and with increasing frequency over the most recent twenty years. Detailed records of channel configurations achieved by successive dredging programmes, as well as quantities of spoil removed, are held by Poole Harbour Commissioners (Engineers' and Hydrographic Departments), who have responsibility for all dredging operations. Figures of total dredged quantities for 1969-1984 are reproduced in McMullen (1982 and 1985) and for March 1985 to April 1990 in Hydraulics Research (1990). For 1969 to 1975 (inclusive) the figure was 138,618m³; 1975-1984, 349,800m³; 1985-1990, 2,017,000m³; 1990-1997, 1,531,336m³ and 1,815,00m³ during the capital dredge between November 2005 and March 2006.The progressive increase reflects the programme of capital dredging over this time period, involving the removal of previously stable sediments to deepen and widen channels to accommodate passage of larger vessels. Material is permanently removed from the harbour to an approved, licensed dredge spoil dumping site offshore and south of Swanage. Some dredging in 1988-9 together with some from the Swash Channel outside of the harbour entrance was used for the renourishment of Bournemouth beach, thus becoming part of the sediment budget of Poole Bay. Approximately 1.1 million m³ of the 2 million m³ material dredged in 2005/6 was used for the replenishment of both Bournemouth and Swanage beaches (Dave Robson, Borough of Poole Council, pers. comm., 2016).

The beach replenishments schemes that have been completed in Poole between Sandbanks and Branksome Dene Chine include: In 2003 using beneficial material from Poole Harbour Commissioners maintenance dredge 88,000m³ of sand was placed between Sandbanks and Shore Road; In 2005 to 2006 using beneficial dredged material from Poole Harbour Commissioners capital dredging to deepen the main shipping channel into Poole Harbour, Poole, Bournemouth and Swanage beaches were replenished with 1,100,000m³ of material; In the winter of 2013/2014, 139,000m³ of beneficial material from Poole Harbour Commissioners maintenance dredge was placed between Sandbanks and Canford Cliffs.

Separate records of maintenance dredging are available for the Main, Middle Ship and certain other channels, and are referred to in Section 3.2 on tidally-induced sediment transport within the harbour. Dredging at the harbour mouth and in the Swash Channel is covered in the unit on Poole Bay.

A continuous program of regular bathymetric surveys is undertaken by Poole Harbour Commissioners (PHC) to monitor the bed levels throughout Poole Harbour.  Bathymetric monitoring has been carried out by the Commissioners for decades and is an integral part of the dredging licence process.  Comparative results are provided to the relevant authorities and included in the Harbour Monitoring Reports published periodically on the PHC website. The Commissioners carried out a major Channel Deepening exercise in 2005/6, following an Environmental Statement.  The predictions from that report have been analysed by Hydraulics Research Wallingford comparing the 2005 levels with the actual measured depths 10 years later. The PHC Maintenance Dredge Protocol details the systematic approach to sediment management within the harbour adopted by the Commissioners. All documents are available on the PHC website at www.phc.co.uk

Each major dredging operation has been preceded by modelling of the likely effects on the hydrodynamics, sedimentation patterns and other environmental characteristics of Poole Harbour as a whole and specific areas potentially vulnerable to change. These are contained in several reports, e.g. Hydraulics Research 1990a and b, 1998; HR Wallingford, 2003.

5. Sediment Stores

5.1 Beaches

There are several sites around the harbour margins where narrow sand, shingle or mixed sand and gravel beaches have developed. Gray (1985) has provided a descriptive inventory of all the major locations, with emphasis on their principal sediment characteristics. Whilst some beaches are clearly associated with recent or contemporary erosion of co-adjacent cliffs, and are affected by littoral drift, others appear to be non-active or relict. The major cause of beach abandonment has been the growth of Spartina saltmarsh and mudflats in foreshore zones, a trend that is now in the process of reversal. In some cases, beaches have developed spit or cuspate elongations, e.g. around lagoons at the north-eastern extremities of Brownsea and Furzey islands; at the western and eastern sides of the channel draining Lychett Bay and along the Arne Peninsula (May, 2005).

There are places where material other than sand or shingle contributes to beach composition, e.g. a small post- Spartina invasion shell chenier at the tip of the main marsh at Keysworth and large cobbles or boulders at several locations. In the latter case derivation is from the erosion or deterioration of seawalls and breakwaters erected up to 120 years ago, composed of Purbeck Limestone from quarries inland of Swanage. On the north shoreline east of Poole town, the natural material is mostly medium-textured sand, which coarsens eastwards. Here, and at many other locations in the central and eastern parts of Poole Harbour, the beach is fronted by extensive, eroded mudflats. Shingle everywhere is poorly-sorted, indicating limited wave action. In a few locations, e.g. north of Hyde's Quay on the south-west of the Arne Peninsula, there has been limited development of low sand dunes derived from fine-textured sediments exposed at low tide across the adjacent foreshore. The distribution of some shell and fine shingle cheniers, at mudflat and marsh margins, may have escaped description.

In a study of beach and spit dynamics at Shipstal Point, where documentary evidence reveals the development of a spit in 1785, May (1976; 2005) indicates that the sediment source is exclusively from adjacent cliff erosion of Bagshot Sands and overlying Plateau Gravels. The beach at Shipstal rests on eroded mudflats, whilst the spit to the north (Photo 4) has retreated landwards (though, overall, the history of this feature illustrates progradation, as shown by overwash gravel ridges dating back to the eighteenth and nineteenth centuries in between marsh sediments behind the present spit). Beach expansion in both northwards and southwards directions occurred following Spartina invasion on Long and Round islands, in Arne Bay and Middle Creek; the latter event modified the local wave climate, promoting sand and gravel deposition. Retreat and erosion has been dominant, estimated at approximately 20m since about 1970 (May, 2005). Gray (1985) observed that the "steeply-shelving" sandy beaches on the western and central sectors of the Arne Peninsula was actively eroding, as was the gravel beach along the southern shoreline of Brownsea Island (Appleton, 1995; May, 1998). The only area of recent beach accretion appears to be Newton Bay and the western coastline of the Goathorn Peninsula. Comparison of air photo cover of the north-east shore between North Haven Point and Poole Head for 1924 to 1982 led Gray (1985) to state that the sand and gravel making up the beach here had been naturally accumulated as there are no records of artificial introduction of sediments to this or any other sites along the harbour shoreline.

Given that there has been little study of beach dynamics, it is impossible to come to firm conclusions concerning harbour-wide trends and patterns. With the exception of the work of May (1976; 2005), the evidence is sketchy and incomplete. However, it may be inferred from field observation that there is little exchange between the coarser material of the beaches and adjacent inter-tidal fine sediment stores. The contribution of beach material to the sediment budget of the harbour is probably small, but full understanding requires research on changes in sediment availability, wave fetches and wind/wave direction frequencies. Green (1940) and Hydraulics Research (1990) note that wave-induced erosion of sand along the north-east shore is very small, though ripple sets are not uncommon (Gray 1985). These may be related to tidal motion, or some combination of wave disturbance and tidal flow. More recent evidence of poorly understood movement of sand occurred during the period 2007 to 2012 where sand at Whitley Lake accreted over the saltmarsh (Dave Robson, Borough of Poole Council, pers. comm., 2016).  

5.2 Sand Banks and Mudflats

Extensive accumulation of unconsolidated clay and silt ("mud") exist around the southern, western and north-western margins of Poole Harbour and represent a long-term net imbalance of the sediment budget of the harbour in favour of accretion. In the north-eastern sector, mudflats form eroded foreshores of limited horizontal development. Within the central and eastern parts there are well-defined banks, or shoals, that appear to be composed predominantly of fine to medium sand, although thin and patchy superficial mud deposits are present. All of these banks are separated by channels, some of which have been dredged over a long period of time for navigation purposes. 80% of the harbour is inter-tidal, dominated by mudflat estimated to be 1,500 ha in area in 2008 (Cope and Mortlock, 2012).

The pattern of the major channels, via which the ebb tide is drained, appears to have been stable at least since 1785 (Gray 1985; McMullen, 1982, 1984 and 1985; Raybould, 1995). However, hydrographic surveys carried out over the past 150 years have revealed detailed changes in the plan form of both banks and channels, presumably resulting from tidally-induced transport. The configuration of the mudflats has changed substantially, primarily due to the entrapment of fine-grained sediments by Spartina anglica (Section 5.3).

The comprehensive hydrographic survey of Poole Harbour (Green, 1940), has been used as a baseline reference for later, partial, surveys and analysis of change (McMullen, 1982; 1985; Hydraulics Research 1990). This source (partly updated in Green, et al., 1952) and the McMullen studies use several nineteenth and early twentieth century charts and plans to identify areas of significant change. Not unexpectedly, work has concentrated on areas of the harbour bed in the vicinity of the principal navigation channels.

The main changes in channel depth, water depth above sandbanks and shoal outline are not given here as they are detailed in the above sources. In general, at high water much of the harbour is covered to a depth of less than 4m. The capital dredge in 2005 increased channel depths to 7.5CD, however depths at Brownsea Roads is 10mCD and at the harbour entrance 10-14m. The average water depth of the harbour is 0.48m (Dave Robson, Borough of Poole Council, pers. comm., 2016). Green (1940) observed that there had been a tendency for the Main Channel to migrate eastwards between 1829 and 1934, especially where its curvature was greatest. He also noted that the Middle Channel had narrowed and deepened over the same timescale. McMullen (1982, 1985) concluded that the main changes between 1938 and 1980 were: (i) the deepening of the Middle Mud area, thus reducing the volume of ebb flow in the Main Channel; (ii) some eastwards and north-eastwards migration, and narrowing, of the Main Channel; (iii) shallowing of the entrance channel between North Haven and Saltern's Beacon; and (iv) stability of the shape and depth of the Upper Wych channel, though it has become a "more prominent" flowline for the ebb tidal stream. These trends continue those that took place earlier in the century (1904-1938), notably the narrowing, deepening, and eastwards shift of the boundaries of all the main channels, the persistence of which is confirmed by Raybould (1997).

The extensive shallowing of the Upper Wareham channel between 1934 and 1954 (McMullen, 1985), the probable result of the release of clay and silt by the decline in Spartina anglica, has continued intermittently up to the present. Raybould (1997) reports that its average depth decreased from 1.7m to 0.75m between 1980 and 1995. Siltation resulting from the breakdown of mud stores in the upper harbour probably accounts for some localised shallowing of the main channels in the 1940s and 1950s. One sandbank, Soldier Ground, disappeared between the late 1930s and early 1950s. Overall, the longer-term trend prior to the spread of Spartina anglica was for slow shoaling and shallowing of harbour channels. This was reversed by the accretion and consolidation of sediment by rhizome growth, amounting to over 1 million tons in the upper harbour alone between about 1900 and 1955 (Hubbard and Stebbings, 1968). Much of this quantity was previously mobile mud, rather than newly-introduced sediment. With the "die back" of Spartina, shallowing experienced in the upper harbour is becoming the net trend elsewhere, thus restoring the expected tendency.

The general pattern of sediment composition of the seabed is simple, with a gradient from silts and clays in the upper harbour to medium to coarse sands, and some superficial patches of gravel, in the area adjacent to the harbour mouth. Only one area of bedrock is - or was - exposed, east of Brownsea Island (Howard and Moore, 1988). There is a general tendency for the banks and flats to be composed of slightly coarser sediments where they form the channel boundaries, and to be finer within the inter-tidal zone. Towards the harbour mouth, coarse sand may be armoured by fine gravel and shells (Hydraulics Research, 1991). Green, et al., (1952) mention the presence of shell fragments "in some areas", although whether these are concentrated is not clear. Green et al., (1952), Dyrynda (1987) and HR Wallingford (2003) remark that the sand fraction in Poole Harbour becomes progressively finer in a westwards direction, but that an embayment of coarser sediment extends northwards across the Middle Ground from north of Brownsea Island. Here, and to the north-east of Brownsea, there is a zone in which there is evident mixing of the main sand size fractions. Royal Haskoning (2005) provide summary details of the distribution of silty and sandy sediments in the main navigation channels and the Turning Basin, reflecting the effects of suspension and redeposition following previous dredging operations.

The significance of these patterns of sediment distribution and composition are discussed in Section 3.2. It must be stressed that information is derived from small samples of sediments obtained during short-term surveys, and that the recovery methods used by Green (1940) and Green et al., (1952) may have significantly underestimated the fine sand and silt fractions. 129 boreholes were taken in the channel as part of the EIA for the capital dredge scheme in 2005 (these are held by Poole Harbour Commissioners) (Dave Robson, Borough of Poole Council, pers. comm., 2016)

5.3 Saltmarsh

Extensive mudflats characterise long sections of the shoreline of Poole Harbour, particularly along the south coast and either side of the Upper Wareham Channel where wave and tidal energy are reduced. The sources of the clay and fine silt particles that have provided the material for mudflat construction include river suspended load, marginal erosion and reworking of deposits on the harbour bed. The processes of accretion are not fully understood, and the initial establishment of the pattern of creeks and channels that subdivide the mudflats into major and minor units has yet to be explained (Raybould, 1997). However, there is no doubt that the rapid colonisation of mudflats by saltmarsh vegetation was a decisive factor in accelerating rates of accretion and consolidation up until the mid-twentieth century. (Humphreys and May, 2005). In this context, the role of the perennial grass, Spartina anglica, over the past century has been critical, although other saltmarsh species and adjacent, though discontinuous, reedbeds have also made a contribution (McInnes, et al., 2011).

The ecological history, taxonomy and genetics of Spartina anglica will not be discussed here; appropriate background is given in Goodman (1958, 1959), Hubbard (1965), Gray and Benham (1990), Gray, et al., (1991); Raybould, et al., (1991); Gray et al., (1995), Raybould (1997) and Humphreys and May (2005). It is sufficient to note that prior to the late 1890s the mudflats in Poole Harbour were primarily colonised by eel grass (Zostera) and algae. Spartina was first reported in 1890, and it thereafter invaded large areas between approximately +0.5mOD and +0.9mOD with remarkable speed and vigour in the succeeding 25 to 30 years (Raybould, 1997). Documentary evidence in the form of large-scale maps, air photos, naturalists' field notes and records, collections of ground-level photographs and published papers has been used to provide a reasonably accurate picture of the pattern of spread of Spartina anglica in the first three decades of this century (Ranwell, 1964; Hubbard, 1965; Raybould, 1997). It did not, however, colonise all areas (such as the north shore east of Holes Bay), and towards the harbour mouth it tended to form isolated clumps. It was probably at its maximum extent in the mid-1920s, covering some 800ha. Starting at a few locations in the late 1920s and early 1930s, Spartina anglica began to show signs of loss of vigour, leading eventually to widespread "dieback". This pattern has been fully described (Hubbard, 1965; Gray and Pearson, 1983; Gray, 1985; Raybould, 1997) and the process of fragmentation and recession continues up to the present. 355 ha were lost between 1924 and 1990- approximately 45% of initial cover (Gray et al., 1991). Research by Born (2005) and Cope et al. (2012) has provided detailed calculation of rates of loss of saltmarsh throughout the harbour, based on georectified historical air photo interpretation and analysis. These studies reveal a reduction of 245 ha between 1947 and 1993 (a 38% reduction), with a progressive decline in rates of loss beginning in the mid or late 1970s. Cope et al. (2012) examined data provided by the Southeast Regional Coastal Monitoring  Programme, 2005-8 and calculated that the contraction of saltmarsh between 1993 and 2008 was only fractionally more than 0.1% of initial area, thus confirming that whilst loss has continued , it has been at a rate significantly below that recorded in the decades between 1940 and 1990. This narrative of ecological change, with its impacts on sedimentation, is a feature of all of the estuaries of the central-southern coast of England (Goodman, 1959; Gray and Benham, 1990); however there is more detailed data for Poole Harbour than most other locations. Gray et al., (1995) have calculated that Spartina expansion captured some 7 million m³ of fine textured sediment, 4 million of which had been released by 1990 following the preceding 60 years of dieback.    

The ecological dynamics of Spartina invasion and recession are complex, and will not be elaborated here (Gray, et al., 1991; Raybould, et al., 1991). As a general point it must be emphasised that the concept of comprehensive spread followed by sudden decline is an unwarranted generalisation. There are some locations where Spartina anglica has maintained its vigour and capacity for colonisation of vacant mudflats, most of them in the upper reaches of Poole Harbour (Hubbard and Stebbing, 1968; Gray, 1985; Raybould, 1997). Other areas have failed to attract Spartina anglica, presumably because of adverse environmental factors. Several researchers have attempted to define critical environmental factors that stimulate or inhibit growth, using both field and laboratory control experimentation (Hubbard, 1968; Goodman, 1959; Ranwell, 1964). These include substrate permeability/porosity; duration of submergence during tidal cycles (cumulative immersion times); sea-level rise and tidal scour; turbulence and light intensity, and reproductive mechanisms. None of these investigations have been conclusive, though there has been a limited consensus in favour of water logging associated with anaerobic soils and sulphide toxicity initiating "dieback". Physical erosion at marsh edges needs to be distinguished from "dieback" proper, which involves patchy degeneration within the body of Spartina swards. Genetic factors may ultimately provide the key to understanding the history and present condition of Spartina anglica in Poole Harbour and elsewhere (Raybould, et al., 1991a and b), as well as the ergot fungus (Claviceps purpurea) (Raybould, 1997).

Following laboratory experiments, Tsuzaki (2004; 2010) found anaerobic soil conditions with impeded drainage to be the most likely cause of the dwarf growth forms and lack of re-colonisation of pans and mudflats by Spartina anglica on the south coast of England. He concludes that the ultimate demise of the Spartina anglica marshes of the south coast of England is the result of frontal and creek erosion of the mature marsh and the failure of Spartina anglica to establish itself on the newly exposed sediments of the foreshore.

Since the 1940s, some areas of Spartina have also been lost through invasion by other species, e.g. Phragmites communis and Scirpus maritimus, from the landward edge (Raybould, 1997). Born (2005) and Cope et al. (2012) calculate that Phragmites reed swamp expanded by 63% of its previous cover between 1947 and 1993 (47ha), with continuing, accelerating colonisation up to 2008- notably in Holes and Lytchett Bays. Between 1972 and 1993 reed swamp increased its habitat by 8.4%, and by a further 17% in the succeeding fifteen years.  Documentary evidence has been used to give various estimates of rates of horizontal expansion of Spartina marshes in Poole Harbour. Hubbard (1965) gives a figure of 15cm of advance during 1964 for the then actively accreting seaward margin of Keysworth Marsh. For the adjacent area of Holton Mere, Hubbard and Stebbing (1968) estimate a rate of 5mm-¹, 1905-1925. Gray (1985) has calculated that Spartina occupied 775 ha of intertidal mudflats in 1924, which had reduced to 694 ha in 1952, and marginally less than 400 ha in 1980 (Gray and Pearson, 1983); the figure of 873 ha for 1962 quoted by Bird and Ranwell (1964) would appear to be anomalous.

Spartina anglica was therefore most extensive in the mid or late 1920s, and has contracted in most areas of the harbour subsequently. Loss has been most rapid in the eastern and southeastern areas, for which Gray (1985) states that 360 ha (46%) of the area colonised by Spartina in 1924 had been lost by 1983, with 189 ha (67% of the area) of the total loss occurring since 1952. "Die-back" first occurred close to channel margins (Oliver, 1925), a process that later spread to the main inter-channel flats through "pan" degeneration. This involved degradation of Spartina anglica around numerous small depressions, with their subsequent coalescence. In areas of advanced dieback, Spartina anglica is now restricted to isolated clumps, e.g. Brands Bay, Holes Bay and the western shoreline of the Arne Peninsula. Most authors have observed that there is an intimate association between Spartina loss and marsh erosion, evidenced by creek erosion and low retreating micro bluffs at the edge of mudflats. Once established, erosion may be a major factor promoting the further reduction of Spartina marsh, but no positive feedback relationship has been proven.

The simplest conclusion is that during the 30-40 years of vigorous expansion, Spartina anglica established itself in vacant habitats that later proved to be unfavourable and in places was dislodged by invaders. As it degenerated, silt and clay that had been stabilised by the plant's deep and tangled root network was re-mobilised through exposure to both tidal and wave scour (Hubbard, 1965; Raybould, 1997). The decline of Spartina is still continuing, but there is a possibility that it may in the future re-invade areas from which it has been displaced (Cope et al., 2012). This would establish an erosion/invasion cycle. Alternatively, Spartina may eventually stabilise, in a lower area of colonisation within the tidal frame (Gray, et al., 1995).

Large quantities of silt and clay have been, and continue to be, temporarily stored in mudflats consolidated by Spartina salt marsh. Ranwell (1964) calculated a rate of 0.5-1.0cm vertical accretion across a series of transects in inner Keysworth Marsh monitored between February and August 1962. Hubbard and Stebbing (1968) derived a much higher figure, 6cm, for January to September 1963, and 14.5cm in 1966 at the same site. For an adjacent site, these authors calculated that Spartina had stabilised some 1.8m of silt and clay, allowing for compaction. Whether these rates, which relate to a small area of vigorous Spartina growth, are characteristic of those that prevailed in the early twentieth century is difficult to assess, although between 70 and 100cm (upper harbour) and 35cm (lower harbour) are possible (Raybould, 1997). However, they are closely comparable to rates of vertical accretion measured for other areas of healthy Spartina on the southern, southwestern and eastern coastlines of England (Hubbard, 1968). Ranwell (1964) reported that silt and clay accretion in Phragmites communis reed swamp, landwards of the Spartina sward, was occurring at the same rate as for the Spartina sites. It is a strong possibility that the contribution of other wetland species to mudflat development has been under-estimated, as Phragmites, Elymus and Halimone have successfully invaded some areas vacated by Spartina since the early 1960s (Gray, 1985, Cope et al., 2012).

The evidence for lateral and vertical accretion rates and spatial change in the extent of Spartina has been used to estimate approximate volumes of silt and clay initially trapped and subsequently released from mudflats (Gray, 1992; Raybould, 1997). The greatest quantities of sediment taken out of this temporary store must derive from areas of erosion open to renewed wave and tidal abrasion. It might be expected that a significant increase in suspended sediment in Poole Harbour has occurred as a result, but there is no convincing evidence for this (Green, 1940). It is possible, indeed probable, that much remobilised clay and silt-sized sediment has either been deposited on banks and shoals (Raybould, 1997) or re-incorporated into accreting mudflats elsewhere. The evidence for continuing deposition on the harbour floor is ambiguous, though the sediment samples recovered by Green (1940) gives indirect proof of this mechanism. This may suggest the concept of an essentially closed system of clay/silt transport, subject to re-working and recycling between sites (Ranwell, 1964; Raybould, 1997). Unless or until quantitative data is available to indicate significant output of fine-grained sediment via the harbour mouth, the only conclusion that can be drawn is that mudflats represent a major component of a large fine-grained sediment store trapped in a slowly subsiding basin. May (2005) estimated that approximately 3 million m³ of sediment continued to be retained in inter-tidal marshes around the turn of this century.

5.4 Volume of Sediment Storage

In a detailed account of the stratigraphy and thickness of sediments at Keysworth Marsh (Hubbard and Stebbing, 1968), documentary records and augured boreholes give a calculation of the increase in marsh height as a result of mud and silt accretion by Spartina anglica. 180cm is the mean value for the marsh as a whole derived from several sample sites, and is based on measurement of sediment thickness above horizons with only Zostera remains (Zostera was the dominant salt marsh plant prior to Spartina). As the documentary evidence indicates that Spartina anglica first appeared at Keysworth in 1910, this gives an accretion rate of over 3cmm-¹, 1910-1960. Considerable inter-annual variation was apparent from the record; allowance was made for post-depositional compaction and settlement, though this factor is not readily calculated. Gray (1992) has proposed that rates of sedimentation during the “peak” of Spartina dieback would be between 6.5 and 1.1mm per year, according to location. Ranwell (1964) gave figures of 35cm of net accretion at Brands Bay and 70cm at Arne, suggesting that there has been considerable spatial variability in accretion rates in recent decades. The number of samples available to Ranwell was restricted and the reliability of his calculations are open to question. However, in general terms, based on understanding of estuary hydrodynamics and ecological factors, it is to be anticipated that rates of accretion in the lower part of Poole Harbour would be less than in the upper sector.

Despite these qualifications, Hubbard and Stebbing (1968) extrapolate from their figures and calculate that Spartina anglica has stimulated mud and silt accretion in the upper part of Poole Harbour to the extent of about 1,000,000 tonnes between 1900 and 1955. Volumes in the middle and lower parts of the harbour would be between 25 to 50% of this total. (See Section 5.3 for other volumetric estimates.)

Although Hubbard and Stebbing (1968) report a maximum total thickness of 6.25m of estuarine sediment accumulation in Keysworth Marsh, there is currently insufficient data to calculate the volume of clay and silt accumulation throughout the harbour since Holocene sea-level rise stabilised at about 3,500 years BP (Edwards, 2001). Borehole logs for land claim and development sites are available, and their detailed correlation would give an improved basis for estimating the dimension of the sediment store in Poole Harbour as a whole (May, 2005). Most of these logs record sedimentation pre-dating accretion due to Spartina colonisation. The problem with this latter phase of accumulation of fine grained sediment  is that spatial variation is likely to be considerable, e.g. Hubbard and Stebbing (1968) report stratigraphical evidence of living Spartina roots to maximum depths of nearly 300cm, on Keyworth Marsh; their figure of 180cm is therefore a generalisation. Using Pinus pollen in a sediment core in the northern area of the harbour, Long et al., (1999) calculate that sediment accretion rates between 1750 and 1890 increased from 0.29 to 1.14mmm-¹ and then to 7.17mmm-¹ during the main phase of the spread of Spartina anglica, 1900-1925. This trend is indirectly endorsed by the results of research into sedimentation in relation to the presence of modern tracers (Cundy and Croudace, 1995). Gray (1992) suggests sediment accretion rates of 6.5mm and 1.1mm per annum when comparing two Spartina marshes at different locations within the harbour.

Further data of relevance concerns changes in the configuration of the harbour channels and creeks. The records held by Poole Harbour Commissioners (McMullen, 1985), and published by Green (1940) and Green, et al., (1952), together with other hydrographic/bathymetric surveys, might provide scope for calculating net volume loss or gain for various locations during the present century. There are, however, major problems in converting planimetric data into volumetric estimates. Permanent outputs due to navigational dredging must also be considered, but records are incomplete and insufficiently site-specific.

5.5 Suspended Sediments

Green (1940) gives some values for suspended silt concentration in the middle and lower parts of the harbour, obtained from laboratory analysis of water samples, which he maps at a small scale. A mean of 120ppm for the Upper Wareham Channel declines to between 60 and 90ppm off Russell Quay and Gigger's Island, and to less than 30ppm in the area near the junction between Main and Middle Channels. Maximum values were consistently obtained at low water, coinciding with minimum salinities. Mineralogical analysis proved that the clay and silt particles were identical to those forming the mudflats in the upper and middle regions of Poole Harbour, leading Green (1940) to presume that the suspended load in the main tidal channels was derived from wave abrasion of mudflat surfaces exposed as tidal levels fell. A further source might have been wave and tidal scour of the eroding margins of Spartina marshes. Green's data was obtained from five north to south transects undertaken in July. A survey carried out during a single day in December (McMullen, 1985), using an attenuated light source siltmeter, revealed very low suspended sediment concentrations in the middle area of the harbour; values are not reported, but were apparently consistent with those given by Green (1940). A report by Hydraulics Research (1990) on the Middle Channel makes the general observation that "mud" concentrations are usually low - at or less than 50ppm - in the lower harbour area. Wave motion is considered to be of some importance in maintaining suspended load, particularly in shallow water where bed stresses are greater. For the area east of Little Channel, in water depths of less than 0.8m, and for at least 50% of the time, mud may be disturbed and suspended without necessarily showing any net transportation. For higher or steeper waves, or deeper water (particularly where these conditions combine), this may not necessarily follow. The field survey undertaken for this investigation was on station during the severe gale of 30 January 1990 when values of up to 600ppm suspended silt concentrations were recorded off Poole Quay and 200ppm at low water near the Middle Channel. The source of this material was stated as erosion "higher up" the western part of Poole Harbour. Whatever the validity of this assertion, the above values are an illustration of the important role of high magnitude, low frequency events that can significantly accelerate both erosion and deposition over short intervals of time. Pethick and Forster (1994) state that highest suspended sediment concentrations occur in the entrance channel, the main channel east of Brownsea Island and other channels between the several small islands in the southern area of the estuary. The basis of their sampling is not clear.   

All of the authors quoted, and others (e.g. Bird and Ranwell, 1964) consider that suspended sediments derived from erosion of mudflats, shoals and marginal cliffs are partly directed to lower energy areas of the harbour, where they eventually settle out (e.g. Lytchett and Holes Bays) and/or are discharged via the harbour mouth at low tide; no authorities, however, provide any quantitative data to support this contention, although a study of Chapman's Peak (Hydraulics Research, 1991) mentions the presence of a "wash load" moving backwards and forwards with reversing tidal streams. The key issue is the net balance between input and output, and this has not been quantitatively evaluated. (See unit on Poole Harbour Entrance to Hengistbury Head for further discussion).

There is a maintenance dredging dispersal site for fine sediment by Brownsea Island within Brownsea Roads, it is licenced to receive up to 30,000m³ per year, currently 15,000-20,000m³ per year is being placed there (Dave Robson, Borough of Poole Council, pers. comm., 2016). Further information may be found at http://www.scopac.org.uk/sediment-sinks.html

6. Summary

1. Poole Harbour is a shallow basin-like estuary, with an irregular, indented coastline of just over 100km. It is a product of postglacial sea-level rise, but its shape and plan form reflect both the original topography of the inundated valleys of the Frome and Piddle rivers and tributaries, together with subsequent modifications by marginal erosion, accretion and land claim.
2. Physical and biotic characteristics show distinct west to east environmental gradients reflecting the increasing marine influence and energy levels in that direction. Artificial reclamation and shoreline protection have modified much of the north and northeastern sectors of the harbour, but elsewhere essentially unmodified natural processes and habitats continue to operate.
3. Extensive mudflats that have accumulated within a gently subsiding structural basin reflect a long-term trend for deposition of fine sediments. Materials are derived from rivers draining into the harbour, primary production of organic material on mudflats and saltmarshes and from the erosion of the harbour margins. Net input of suspended and perhaps coarser sediments from marine sources, via the harbour mouth, is probable but has not yet been satisfactorily demonstrated. Overall inputs would appear to be relatively slow, so it is proposed that the harbour operates naturally as a relatively closed circulation system, subject to redistribution of existing accumulated sediments.
4. Although a tendency is identified for net seaward tidal flushing of bedload sediments (mostly sands) out of the harbour, the extent to which this actually occurs remains uncertain. It could be that sediments become stored in circulation systems within the harbour, or that outputs are offset by periodic wave driven inputs through the entrance. The transport pathways and rates of movement of bedload and suspended load sediments are not sufficiently well understood to permit a reliable overall assessment.
5. Reclamation and channel dredging have impounded, directly removed or redistributed large quantities of sediments from and within the transport and storage sub- systems. The former practice has reduced the harbour area diminishing its tidal prism and potentially affecting tidal currents. Full and longer-term effects of these impacts on the sediment budget have not been determined. This is being mitigated against by the dispersal site for fine sediments.
6. The most significant events over the past century have been associated with the expansion (1890s – mid-1920s) and subsequent dieback (1930s to present) of Spartina anglica dominated mid and lower saltmarsh. Colonisation induced rapid accretion raising the marsh surface by up to 1.75m and covering a maximum area of 800 hectares in 1924. Dieback resulted in the erosion of the marshes back down to low mudflats and the release of the accumulated sediments leaving less than 400 hectares remaining at present (Born, 2005; Cope et al., 2012). The fate of the muds released remains uncertain, although it is postulated that they could have contributed to growth of mudflats and shoals at new locations as well as causing siltation of channels. In addition, reedbed habitat has continued to increase in area since 1993 (Born, 2005; Cope et al., 2012).
7. In response to the sequential growth and dieback of extensive Spartina marshes, harbour shorelines first became protected from wave action and are now becoming increasingly exposed, resulting in reactivation of inactive clifflines and growth of marginal beaches and spits as sands and gravels released from cliffs drift slowly along shorelines.
8. The low gradient intertidal shorelines and hinterland of the south and western harbour margins are identified as being especially sensitive to the effects of future climate change and sea-level rise. Defended portions face the prospect of increasing “coastal squeeze” and diminishing intertidal foreshore and habitats (Halcrow Maritime et al., 2001; Cope et al., 2012). Future management policy will need to consider the possibility of planning for controlled tidal inundation of some low-lying hinterland areas to create intertidal habitats through strategic policies of both non-intervention and managed realignment.

7. Opportunities for Calculation and Testing of Littoral Drift Volumes

Data collected by the Defra-funded National Network of Regional Coastal Monitoring Programmes is pivotal for future improvement in estimating beach change.  The Programmes consist of topographic beach surveys, nearshore bathymetry, aerial photography, lidar, coastal hydrodynamics (waves and tides) and terrestrial habitat mapping.  Specifications for data collection are consistent for all regional programmes and the data and analysis reports are made freely available under the Open Government Licence from www.channelcoast.org.

The Southeast Regional Coastal Monitoring Programme commenced in 2002. The Lead Authority is New Forest District Council, with data collection, analysis and reporting led by specialist teams at the Channel Coastal Observatory (CCO), Canterbury City Council and Adur and Worthing Councils. Longer-term Coastal Monitoring Programme data, when combined with other data sets, academic research and historical studies may enable sediment budgets, transport rates and directions to be identified and/or validated in the future.

No specific research has been undertaken on the numerical estimation of rates and volumes of littoral transport in the harbour. In view of its limited beaches, discontinuous set of pathways, and the small quantities and rates involved, it is not considered a priority. Instead, research efforts need to focus on gaining a first approximation of the harbour's sediment budget.

8. Research and Monitoring Requirements

The Southeast Regional Coastal Monitoring Programme commenced in 2002. Analysis of the data between 2006 and 2012/13, combined with other datasets, academic research and historical studies has enabled sediment budgets, transport rates and directions to be identified or verified. However, at certain sites either due to a lack of long-term data, data coverage or sedimentological information (e.g. composition and proportion of beach grade material arising from cliff erosion), quantification of sediment transport rates of gravel and sand has not been possible.

The following important questions remain uncertain:

1. Is the overall sediment budget positive or negative at present and is there a net input or output at the harbour entrance?
2. What are the detailed pathways of bedload and suspended load transport?
3. What has been the fate of the large quantities of fine sediment that have been released following the large-scale die-back of Spartina since the 1930s?
4. Is the morphological configuration of the harbour and its channels close to a state of equilibrium? Is reclamation and dredging likely to have affected this relationship?
5. Is the dispersal site at Brownsea Island effective.

Notwithstanding results from the Southeast Regional Coastal Monitoring Programme, and the summarised information collated in the Hurst Spit to Durlston Head SMP2 (Royal Haskoning, 2010b), recommendations for future research and monitoring that might be required to inform management are as follows:

1. The literature review for this account of sedimentation and sediment transport in Poole Harbour is derived from: (i) Southeast Regional Coastal Monitoring Programme data and analysis; (ii) analytical papers based on field and/or laboratory process measurement and monitoring, documentary studies etc.; (iii) general review or "inventory"-style papers, containing useful descriptive material, and (iv) consultants' reports, giving data sets obtained from short-term surveys of bathymetry, waves and tidal currents, sediment composition and sediment transport processes for specific areas. The main limitation of material in this last category is its temporal and spatial restrictions, particularly in a large estuary where there are constant reversals in tidal stream and sediment transport vectors. The main question concerns the representativeness of the short empirical data sets collected for specific areas, less its reliability.
2. It has therefore proved impossible to construct a full qualitative model of the sediment budget of the harbour. Some budget elements and transfers (losses and gains of sediments) are reasonably well established, but there are many ambiguities and areas of uncertainty. A critical omission from present knowledge is that of rates and volumes of sediment transport at the harbour entrance. Whether there is a net input or output needs to be established, together with better definition of the stability of sand banks and shoals in the eastern part of the harbour. Most survey work has concentrated on the main navigation channels and adjacent banks, including areas proposed for newly dredged channels. There remain areas of Poole Harbour for which there is very little information on channel and bank stability, particularly in the south and southeast. The dynamics of tidal streams, especially their capacity to abrade sandy surfaces and move sediments by suspension or traction need to be comprehensively studied over representative time periods, i.e. for different tidal cycles and seasons. The contribution of cliff, marsh edge and possibly beach erosion has been given little attention and the intuitive assumption that none are a significant source of sediment load has not been quantitatively explored. The quantity and quality of knowledge of the input of fluvial sediment is also limited. Not until all of these areas of uncertainty have been thoroughly researched will it be possible to confirm the concept of Poole Harbour as a sedimentary sink.
3. An important step is to undertake studies that consider the harbour as a whole rather than to limit attention to small parts. This could involve a co-ordinated comprehensive multi-disciplinary approach, with the central objective of providing better definition of the sediment budget. As a virtually landlocked basin, it should be possible to isolate and quantify the major inputs, stores, throughputs and outputs of sediment. This task, though challenging, is relatively more straightforward than for any part of the adjacent 'open' coastline. A useful initial study would be to calculate the tidal prism for the harbour as a whole and for its main channels at intervals. This could then be compared with the relevant channel cross-section areas to determine the extent to which the two had achieved equilibrium and thus the potential for future morphological changes. This type of approach was used extensively to study the Essex and Solent estuaries within their respective CHaMP studies (Cottle, Pethick and Dalton, 2002; Bray and Cottle, 2003).
4. Sediment sampling coupled with hydrographic survey analysis is needed covering the whole harbour to establish reliable baselines against which future changes might be compared (Colenutt, 1994). It is recommended that analyses of particle size and sorting be undertaken to derive bed transport vectors as was achieved for Chichester Harbour entrance by Geosea Consulting Ltd (1999). It would be useful also to map bedforms throughout the main channels and banks of the harbour as they could provide additional evidence of directions of sediment transport.
5. A baseline of habitat extents was mapped from aerial photography flown in 2008 by the South-east Regional Coastal Monitoring Programme. The aerial photography in 2013 will be remapped. The saltmarsh extents can be added to the historical analysis undertaken by Born (2005) and Cope et al., (2012).This allows projection of likely future changes indicating where future management or habitat creation might be required.