Historical morphological change in the Mersey Estuary

Historical morphological change in the Mersey Estuary

Continental Shelf Research 22 (2002) 1775–1794 Historical morphological change in the Mersey Estuary C.G. Thomasa,*, J.R. Spearmanb, M.J. Turnbullb a...

663KB Sizes 0 Downloads 17 Views

Continental Shelf Research 22 (2002) 1775–1794

Historical morphological change in the Mersey Estuary C.G. Thomasa,*, J.R. Spearmanb, M.J. Turnbullb a

Centre for Civil Engineering, Oxford Brookes University, Gipsy Lane, Oxford OX3 OBP, UK b HR Wallingford, Howbery Park, Oxfordshire OX10 8BA, UK Received 5 December 2000; accepted 3 December 2001

Abstract Several techniques including analysis of bathymetric data, calculation of analytical parameters and computational hydrodynamic simulations are combined in this study to develop a conceptual understanding of processes causing morphological change in the Mersey Estuary between 1871 and 1997. Volumetric analysis demonstrates that morphological change is dominated by a trend of significant accretion between 1906 and 1977, with estuary volume reducing by E10% (70 Mm3), followed by a relatively small increase in volume between 1977 and 1997. Previous research identified the construction of training walls between 1906 and 1936 to stabilise the position of the low water channel in Liverpool Bay outside the estuary as a probable cause of perturbation. The paper examines the hypothesis that sedimentation in the estuary was controlled by changes to hydrodynamic flow and related sediment transport patterns outside the estuary resulting from training wall construction, and that the estuary has now evolved towards a stable state. The results from computational hydrodynamic models for the years 1906, 1936 and 1977 quantifying potential changes in sediment transport pathways from outside the estuary indicate a significant increase in potential sediment supply to the mouth of the estuary during the period of peak accretion. However, these changes cannot be solely attributed to construction of the training walls, but result from the combined effect of training wall construction and dredging activity in the sea approach channels. Furthermore, it is not simply changes in hydrodynamic flow characteristics that cause sedimentation but also the existence of salinity-induced gravitational circulation within the estuary and wider Liverpool Bay system that acts as an important mechanism for importing sediment into the estuary. Evidence for evolution towards a stable estuary state is provided by derivation of a sediment budget demonstrating a negligible net flux of sediment into the estuary between 1977 and 1997. The establishment of a stable state is attributed to a reduction in the calculated transport of sediment across Liverpool Bay reducing the supply of sediment to the estuary mouth. r 2002 Published by Elsevier Science Ltd. Keywords: Bathymetric data; Morphology; Sediment transport; Non-cohesive sediment; Mersey Estuary

1. Introduction The evolution of estuary morphology over periods of E100 years results from changes in sediment transport patterns determined by forcing *Corresponding author. Fax: +44-1491-832233. E-mail address: [email protected] (C.G. Thomas).

factors operating at a range of temporal and spatial scales. A theoretical equilibrium state in an estuary requires that forcing factors produce a long-term averaged sediment flux through the tidal inlet approximating to zero. To achieve morphodynamic stability, i.e. the ability to adapt to longterm changes in forcing factors, Dronkers (1998) suggests there must be a change in physical

0278-4343/02/$ - see front matter r 2002 Published by Elsevier Science Ltd. PII: S 0 2 7 8 - 4 3 4 3 ( 0 2 ) 0 0 0 3 7 - 7

1776

C.G. Thomas et al. / Continental Shelf Research 22 (2002) 1775–1794

processes, particularly tidal propagation, in response to perturbation such that the average ebb and flood sediment fluxes become unbalanced and restore an equilibrium state. Several studies (Fitzgerald et al., 1976; Friedrichs and Aubrey, 1994; Dronkers, 1998) have shown that changes in system geometry alter tidal propagation possibly altering tidal asymmetry and influencing net gain or loss of sediment to the system. However, interaction between the estuary and the seaward environment also influences morphology by determining sediment supply to the system, which has been found to be significant in the Gironde estuary, France (Castaing and Allen, 1981). In a sediment supply limited situation an estuary may exist in a theoretical equilibrium state where ebb and flood sediment fluxes are both very small, but tidal propagation is such that the estuary has the potential to import or export sediment if it were available. In this situation the switching on and off of sediment supply to the system acts as a fundamental control on morphodynamic evolution. The causes of morphological change in the Mersey Estuary between 1911 and 1957 were examined by Price and Kendrick (1963) employing two physical models. The first model represented Liverpool Bay configured to bathymetries for 1911 and 1957 with freshwater flow only, and the second represented the Mersey Estuary upstream of New Brighton including salinity effects. The physical model results demonstrated significant changes in hydrodynamic flow regime in Liverpool Bay following construction of a training wall between 1906 and 1936 to stabilise the position of the low water channel for navigation. The changes in hydrodynamic flow conditions have significant implications for sediment transport pathways, and suggest that estuary morphological change has resulted from a change in sediment supply to the mouth. Based on results from the Mersey Estuary physical model the mechanism proposed by Price and Kendrick (1963) for the transport of sediment through the Narrows to enable sediment to be imported from Liverpool Bay was via density-induced gravitational circulation. The objective of this study was to examine the hypothesis that sedimentation in the estuary was

controlled by changes to hydrodynamic flow and related sediment transport patterns outside the estuary resulting from training wall construction, and that the estuary has now evolved towards a stable state. The study develops previous findings by examining the morphological response of the estuary to perturbation which has led to a situation where the estuary appears to have stopped importing sediment. Sediment supply and net tidal transport patterns the mouth of the estuary are examined to assess their relative influence upon morphological change and recovery. The objective is addressed by examining hydrographic data from the Mersey Estuary recorded between 1871 and 1997. Historical changes in bathymetry are analysed in detail to examine potential changes in flow regime associated with changes in system geometry characteristics. Changes in estuary hydrodynamics are examined in greater detail by applying 2d and 3d computational hydraulic models to bathymetric configurations for the estuary and offshore area for 1906, 1936 and 1977. Finally, hydrodynamic flow model results are employed as a basis for examination of broader flow interactions and associated sediment transport pathways between the estuary and the adjacent Irish Sea.

2. Study area The Mersey Estuary (see Fig. 1) is a strongly tidal estuary located on the West coast of the UK. Elevation ranges at the mouth vary from 4 to 10 m over the extremes of the Neap–Spring cycle. The wave climate in Liverpool Bay is dominated by waves from the West and North-West corresponding to the direction of the maximum fetch length. Frequent South-Westerly winds are also experienced but are limited by a fetch length of 80 km. Data collected at the Mersey Bar light vessel between September 1965 and September 1966 indicated that a significant wave height >2 m had an exceedance of E10% in winter and autumn and E2% in spring and summer (Draper and Blakey, 1969). The Narrows is an inerodible geological constriction at the mouth of the estuary. Approxi-

C.G. Thomas et al. / Continental Shelf Research 22 (2002) 1775–1794

1777

Fig. 1. The Mersey Estuary and Liverpool Bay.

mately 1 km wide with a mean depth of 15 m, tidal currents through the Narrows can exceed 2 m/s. Freshwater flow into the Mersey varies from 25 to 200 m3/s, an order of magnitude less than tidal flow. The estuary exhibits characteristics of a partially mixed estuary in the Narrows, with differences between extreme values of salinity at any position during a tidal cycle of E4 g/l for low river discharge, and 11 g/l for high river discharge (Dyer, 1997). Sediment properties in the estuary differ widely from the inner estuary to Liverpool Bay. The inner estuary comprises extensive intertidal banks of mud and sand. It was observed in a Water Pollution Research Board Technical Paper (1938) that mudbanks are generally located adjacent to the shoreline, whilst sandbanks line the low water channel. High flow speeds through the Narrows scour the bed down to rock and gravel whilst Liverpool Bay contains large areas of sandbanks exposed at low water, with a main navigable low water channel that shifted position until training wall construction between 1906 and 1936 (Wright et al., 1971) found that within Liverpool Bay the seabed surface consists of fine

to medium sized sand formed by tidal current reworking of Pleistocene glacial and fluvioglacial deposits overlying a partly eroded surface of boulder clay. In addition, local outcrops of gravel filling depressions have been found in the boulder clay surface (Sly, 1966). Several major civil engineering works have been undertaken in the estuary and Liverpool Bay area that may have contributed to changes in estuary morphology. Within Liverpool Bay these included: *

*

*

Dredging of the bar at the seaward end of Queens Channel beginning in 1890 to a depth of 6.4 m Liverpool Bay Datum (LBD) in creased to 9 m LBD by 1895. By 1908 1 Mt of material was being removed from the bar annually. Dredging of the sea approach channels, 10 Mt was removed annually in 1908, increased to a maximum of 17 Mt between 1910 and 1917. Construction of training walls (see Fig. 1), commencing with a 3.6 km length on the outside of Crosby Channel bend, extended to the west and augmented between 1914 and 1935 by training walls on both the east and west sides of

1778

C.G. Thomas et al. / Continental Shelf Research 22 (2002) 1775–1794

Crosby Channel. Between 1945 and 57 the training walls were extended in a seawards direction. Within the estuary anthropogenic activity included: *

*

*

*

*

*

Construction of piers for the Runcorn Railway Bridge, completed in 1865; Construction of the Manchester Ship Canal with its associated reclamation of river and tidal water, completed in 1894; Diversion of the River Weaver, completed in 1896; Tipping of slag to form an embankment on the north-east side of the estuary between Runcorn and Hale Head, completed in 1896; Construction of piers for the Runcorn transporter bridge, completed in 1902; Dredging of estuary channels carried out intermittently from the 1890s and culminating in the average annual extraction of 2.75  10 Mm3 from the Eastham Channel during the latter part of the 1950s.

3. Data coverage Data sets adequately covering a period of morphological change in sufficient detail, or with potential to derive processes responsible for change are rare. Historical bathymetric data of the Mersey Estuary employed in this study covering a period of E100 years of substantial morphological change provide a significant resource for examining forcing factors and estuary response to perturbation. The principal data source employed in this study comprises a series of Mersey Conservancy hydrographic surveys of the inner estuary from New Brighton to the tidal limit recorded by the Mersey Dock and Harbour Company for the years 1871, 1906, 1936, 1956, 1977 and 1997. The datasets cover a substantial intertidal area of the estuary, up to 10 m above LBD. The resolution of data coverage is relatively coarse and consists of 174 survey lines, which are repeated for each survey and the spacing of depth measurements along the survey lines varies be-

tween years. In addition, surveys were also employed for Liverpool Bay, covering a square area from the eastern point at the mouth of the Dee to Formby. An example of the spatial distribution of combined bathymetric data for the Mersey Estuary and Liverpool Bay is illustrated in Figs. 2 and 3.

4. Changes in morphological properties 4.1. Volumetric change Volumetric trends were calculated for the inner estuary between New Brighton and Warrington Bridge from spatially interpolated grid based Digital Elevation Models (DEM). Volumetric change in an estuary represents a simple approach to characterise net morphological behaviour, i.e. whether an estuary is accreting or eroding, an increase in estuary capacity represents a lowering of bed levels and less sediment in the estuary and a decrease in estuary volume indicates sedimentation. Between 1906 and 1977 estuary volume decreased representing net accretion followed by a small increase in estuary volume, representing a net erosional trend, between 1977 and 1997 (see Figs. 4 and 5). Estuary volume decreased by E70 Mm3 between 1906 and 1977 representing a reduction in estuary capacity of E10%. Volume calculations are dependent upon the accuracy of bathymetric data. The random error of depth measurements from the 19th century are likely to exhibit significantly greater error margins than present day measurements. However, when averaged over depth measurements covering the whole of the estuary system, random error has a significantly reduced effect upon volume calculations. Systematic measurement errors are also likely to exist in data, particularly as nautical maps were undertaken for navigational purposes and were primarily interested in recording the lowest points. In comparisons of estuary volume, calculation of trends is only distorted in situations where systematic measurement errors alter between surveys, for example due to a change in the means of relating tidal level to depth measurements or change in measurement device. Overall

C.G. Thomas et al. / Continental Shelf Research 22 (2002) 1775–1794

Fig. 2. Principal bathymetric features of Liverpool Bay and the Mersey Estuary.

Fig. 3. Spatial distribution of bathymetric data measurement points in Liverpool Bay and the Mersey Estuary.

1779

C.G. Thomas et al. / Continental Shelf Research 22 (2002) 1775–1794

1780

Fig. 4. Mersey Estuary changes in total water volume (below 9.95 m LBD) 1871–1997.

the change in estuary volume between 1906 and 1977 is significantly greater than would be expected to arise from errors in measurement. Although volume change between 1977 and 1997 is much smaller and would be distorted to a greater degree by measurement error, these two surveys were both undertaken using modern echo-sounding equipment to measure the same survey lines which should reduce the potential for significant differences in systematic error between the two surveys. Volume change may also be affected by sea level rise. In the case of the Mersey a study of long-term change in sea level by Woodworth (1999) found that sea level has risen E1 mm/year over the duration of the last century. Over the course of the last century this amounts to a total of 1 cm which is unlikely to have a significant effect upon volume calculations. 4.2. Sediment budget

Fig. 5. Dredging in the Mersey Estuary above New Brighton 1949–1999 from Mersey Conservator Annual Reports to the Secretary of State.

The value of bathymetric data in the Mersey has been significantly augmented by the addition of data to derive a sediment budget (see Table 1). The calculated budget is not fully comprehensive, with difficulties establishing fluvial input to the Mersey Estuary as records of changes in fluvial flow are disjoint. Fluvial flow was assumed to be negligible as it likely to be several orders of magnitude less than marine sediment exchange. Dredging records

Table 1 Sediment budget for the Mersey Estuary Period

Total water volume change in the Mersey Estuary (Mm3)a

Volume change due to reclamation (Mm3)a

Material dredged from Mersey Estuary (Mm3)b

Disposal of dredged material within Mersey (Mm3)b

Net annual sediment flux (Mm3)c

1871–1906 1906–1936 1936–1956 1956–1977 1977–1997

+23.3 30.2 24.4 14.5 +11.2

6.8 4.5 6.4 2.3 0.0

After 1890 9.3 39.8 19.9 29.3 8.8

n/a n/a n/a 0.2 1.5

0.59 +2.18 +1.75 +1.97 0.20

a

Totals derived from digitising bathymetric data. From Water Pollution Research (1938) and Mersey Conservator Annual Reports to Secretary of State (1939–1979), using a rough approximation of 0.79 m3=1 cubic yard2.7 hopper tons. Where no data is available for quantity of dredged material deposited in the estuary it is assumed to be zero. b

C.G. Thomas et al. / Continental Shelf Research 22 (2002) 1775–1794

also need to be treated with caution as measurements are unreliable and methods of converting hopper tons to m3 are an approximation. In addition, there is no data for the deposition of dredged material within the estuary for the period 1871–1956. Where no data was available deposition in the estuary was assumed to be negligible. Despite data inaccuracies and assumptions, however, a sediment budget indicates net trends in the estuary with greater clarity and is a useful tool in morphological analysis. A sediment budget is of particular value in attempting to establish the stability or otherwise of an estuary system, indicating net flux of sediment into the estuary. The key features indicated by the sediment budget are a high net sediment flux into the estuary between 1906 and 1977 indicating a net import of sediment into the system. This is followed by a relatively small net sediment flux into the estuary indicating that the estuary has established an equilibrium state in terms of net flux of sediment through the estuary mouth, although volume change still occurred as a result of dredging activity. The interpretation of these trends leads to the conclusion that although dredging within the estuary has been ongoing, it has been largely overridden by the effects of a greater net flux of sediment into the estuary through the period 1906–1977. Intensive dredging activity within the estuary over short periods has, however, proven sufficient to reverse the net trend of accretion, such as an increase in estuary volume between 1953 and 1954 in response to high levels of dredging in Eastham Channel (Price and Kendrick, 1963). As net flux of sediment into the estuary has declined dredging within the estuary has exerted greater influence upon morphological change in the estuary and now appears to be the dominant influence upon net morphological change. Thus, an ordering of impacts may be defined with dredging within the estuary largely a second-order effect whilst estuary behaviour was dominated by sediment import through the estuary mouth, but has become a dominant impact as the flux of sediment into the estuary has reduced. Although estuary volume may now be increasing as a result of sediment removal via dredging, this only represents the net estuary trend

1781

with localised patterns of erosion and accretion within the estuary maintaining the requirement for dredging.

4.3. Integrated geometrical properties Changes in geometrical characteristics associated with volume change can provide an indication of interaction between estuary morphology and physical processes. Several studies (Fitzgerald et al., 1976; Friedrichs and Aubrey, 1988; Van Dongeren and De Vriend, 1994) demonstrate that if cross sectional area is widened such that intertidal area is decreased, then flood duration is shortened with respect to ebb duration. Conversely, if the cross section is deepened then the inverse occurs. Through examination of trends in integrated geometrical characteristics of the estuary (Table 2) it is clear there have been significant changes in the system geometry of the estuary, with implications for hydrodynamic and sediment regime in the estuary. Volume changes indicate a linear trend in the subtidal volume, with a significant reduction in volume and associated changes in subtidal channel width and depth. Intertidal volume, however, demonstrates a decrease until 1956 and increase from this point onwards, possibly indicating an initial response and subsequent recovery from perturbation. 1906 and 1936 configurations show a deeper subtidal channel, indicating increased ebb dominance for these periods, which is not compa-

Table 2 Changes in integrated geometrical parameters in the Mersey Estuary upstream of New Brighton Year

1871 1906 1936 1956 1977 1997

Integrated geometrical parameters h (m)

Subtidal area (Mm2)

Intertidal area (Mm2)

Subtidal volume (Mm3)

Intertidal volume (Mm3)

5.7 6.1 6.1 6.0 5.8 5.6

18.5 20.6 19.4 19.8 16.9 16.5

65.7 61.1 60.9 61.5 60.6 60.9

129 143 136 131 123 123

338 328 317 299 317 329

C.G. Thomas et al. / Continental Shelf Research 22 (2002) 1775–1794

1782

tible with significant accretion. However, subtidal area was relatively large in 1906, indicative of a wider estuary and increased flood dominance. Subtidal area demonstrated a significant reduction between 1906 and 1977, reflecting a decrease in the width of the subtidal channel indicative of an increasing ebb dominance in the system through time, reconcilable with observed patterns of volume change in the estuary. The most probable cause of a reduction in estuary width is accretion of sediment imported into the estuary upon intertidal areas. Deposition of dredged material has not contributed substantially to a reduction in estuary width as the majority of material dredged from the estuary was deposited in Liverpool Bay outside the estuary system, and within the estuary material was deposited in the deep water area of the Middle Deep (Price and Kendrick, 1963). In order to achieve morphodynamic stability a change in subtidal depth must be accompanied by a change in subtidal width, if one remains constant then change to the other is an ongoing process and a state of equilibrium cannot be attained. Of critical importance to potential morphological evolution is the balance between the rate of changes in width and depth. If the balance is towards a relatively greater decrease in depth, hydrodynamics should shift toward flood dominance, however, if there is a relatively greater decrease in width, hydrodynamics should shift toward ebb dominance. Studies have attempted to characterise the behaviour of estuary systems by relating geometric parameters to theoretical tidal asymmetry. An analytical approach proposed by Dronkers (1998) based upon parameterisation of 1d tidal equations is particularly applicable to the current study g¼

2 HHW SLW ; 2 HLW SHW

ð1Þ

where HHW and HLW are the mean depth over the estuary at mean high and low water, respectively, and SHW and SLW are the wet surface area of the estuary at mean high and low water. One equates to an approximate balance between ebb and flood tides, figures o1 indicate a dominance of ebb tide over flood, and figures >1 indicate a dominance of flood tide over ebb. The g parameters have been

Table 3 Historical changes in Dronkers g parameter in the Mersey Estuary upstream of New Brighton Year

Morphological g parameter

1871 1906 1936 1956 1977 1997

1.41 1.58 1.53 1.53 1.34 1.31

calculated for the Mersey Estuary system upstream of New Brighton for each of the years with bathymetric data available with values given in Table 3. The data in Table 3 shows a reasonable degree of correlation with expected trends in hydrodynamic regime through the estuary, flood dominance is greatest in 1906, reflecting greatest capacity to import sediment and decreases significantly to 1997, reflecting a reduced capacity to import sediment.

5. Hydrodynamic flow modelling The availability of historical hydrographic surveys enables quantitative comparison of hydrodynamic regime by applying computational hydrodynamic models to historical estuary bathymetric configurations. Of particular importance is examination of the changes in tidal flow patterns through the Narrows as it forms the connection between Liverpool Bay and the Mersey Estuary. The tide at the estuary mouth influences tidal asymmetry in the estuary itself, and tidal flow through the Narrows exerts a significant control upon sedimentation in the estuary. Changes in flow velocity are most likely to be caused by changes in tidal forcing as a result of bathymetric changes in Liverpool Bay as the Narrows is nonerodible and therefore unlikely to have adapted to changes in flow conditions. Due to a scarcity of observed historical data on physical processes for direct analysis, this provides the most suitable means for examining changes in flow conditions. Model simulations employing 1906, 1937 and 1977 bathymetric data for the estuary and Liverpool

C.G. Thomas et al. / Continental Shelf Research 22 (2002) 1775–1794

Bay areas were chosen because of the availability of suitable surveys and significant change occurring between each of these times. Study of the possible mechanisms causing accretion in the estuary is extended by studying changes in sediment transport patterns in Liverpool Bay and the estuary diagnostically using 2d and 3d computational hydrodynamic simulation results. 5.1. The model To model estuary hydrodynamics, shallow water equations were solved using a finite element method employing TELEMAC software (Hervouet and Van Haren, 1994). Identical calibration factors were applied to the flow model representing each bathymetric configuration so differences in calculated flow conditions result from system geometry alone. The model employs a finite element mesh extending to a boundary E40 km offshore. Flow conditions are initially represented two dimensionally as depth-averaged flow conditions, and then represented three dimensionally, resolving full horizontal and depth variation of flow. Both 2d and 3d modelling approaches employed an unstructured mesh, and were set up identically using the same grid. The model covered an area from Fleetwood in the North to Llandudno in the west and includes both the Mersey and Dee estuaries. The model resolution was 100–300 m in the Mersey Estuary and 200–400 m in Liverpool Bay, increasing to 6 km at the offshore boundaries. The 3d model incorporated five vertical layers, equally spaced throughout the water column. 5.2. Initial and boundary conditions No salinity was prescribed in the depth-averaged model simulation. The initial salinity field for the 3d model was vertically uniform, and was prescribed according to data derived from several sources (Heaps and Jones, 1977; Winters, 1984; Ramster, 1975; Water Pollution Research, 1938). No freshwater flow was prescribed for either model; the 3d model was dependent upon the provision of freshwater given in the initial conditions to drive density-induced circulation patterns.

1783

The model was run for a sufficiently short period that there was no significant decay in stratification in the system although salinity was slightly redistributed during simulation. The surface momentum flux (wind stress) is prescribed as zero for all model simulations. The models were run for two spring tides using a repeated cycle starting from high water, and the elevation along the boundary for a spring tide was derived from M2+S2 tidal harmonic constituents. The same boundary conditions were applied to both the 2d and 3d models. Calibration was undertaken for the 1977 bathymetry by comparing model predictions of current patterns in Liverpool Bay with Admiralty Tidal Diamond Data and by comparing model predictions of water level with measurements taken from West (1980). The models reproduced these observations well. A constant Nikuradse friction parameter of 0.005 was employed throughout the model domain. 5.3. Changes in hydrodynamic flow in the Narrows Results are presented graphically for a location (Point A, see Fig. 1) in the middle of the Narrows area of the estuary, which is examined for trends in velocity relevant to potential changes in sediment flux into the estuary. The same location is examined for the 2d depth integrated and 3d bed layer results. Significant differences in flow characteristics can occur across the estuary due to flood and ebb channels. However, comparison with other nearby locations demonstrated limited variation across the cross section for this location. Significantly, the results presented are representative of the relative change in flow characteristics which is a general feature of the hydrodynamics, not just a localised feature. Although sediment transport may be determined by a number of factors, velocity exerts a key influence under normal tidal conditions and velocity profiles through a tidal cycle provide a significant indication of possible changes in sediment transport regime. Analysis of 2d results (see Fig. 6) at Point A in the Narrows (see Fig. 1) shows an increase in tidal current speeds on both the ebb and flood tides between 1906 and 1936. There is a larger increase

1784

C.G. Thomas et al. / Continental Shelf Research 22 (2002) 1775–1794

Fig. 6. Change in flow velocities in the Narrows at Point A from 2d modelling results.

in ebb velocity reflecting a decrease in flood dominance, i.e. an increase in ebb velocity relative to flood velocity potentially reducing transport of sediment into the estuary. However, flood velocity is still larger than ebb velocity in this location, and since flow velocity is non-linearly related to sediment transport, considerable potential still exists in 1936 for the estuary to import sediment via hydrodynamic flow conditions. The calculated changes between 1936 and 1977 demonstrate a decrease in tidal current speeds on both the ebb and flood tide. The decrease of flood velocity is markedly greater than ebb velocity and the two velocities become relatively equal in 1977, exhibiting a greater balance between ebb and flood velocity than other years studied. The indication from this result is that the estuary in 1977 has reduced potential to import sediment, this being reconcilable with the establishment of an equilibrium state with negligible change in estuary volume.

Fig. 7. Change in flow velocities in the Narrows at Point A from 3d modelling results.

Analysis of 3d bed layer results (see Fig. 7) at the same location, Point A reflects changes in 2d depth-averaged flow velocities. Tidal current speeds increase by approximately the same margin on both ebb and flood tides between 1906 and 1936. Flood velocity remains larger than ebb velocity, demonstrating considerable potential for the estuary to import sediment via hydrodynamic flow conditions. Changes between 1936 and 1977 demonstrate a decrease in tidal current speeds on both the ebb and flood tide, to a level similar to 1906. The two velocities become more equal in 1977, exhibiting a greater balance between ebb and flood velocity than other years studied, although flood velocity is larger with respect to ebb velocity than for the 2d depth-averaged results. The similar trends exhibited by 2d and 3d results has important implications for the physical processes controlling change in the estuary, indicating that density driven circulation, represented only in the 3d model, has been of limited importance to the net movement of sediment through the Narrows as

C.G. Thomas et al. / Continental Shelf Research 22 (2002) 1775–1794

both models produce comparable flow calculations. The 3d flow velocities are of significant importance to sediment transport as they resolve lateral depth variations with current depth structure, and are well suited to modelling of coarser sediments sensitive to near bed current profiles. 5.4. Changes in potential sediment transport Anecdotal evidence (Kendrick, pers. comm.) from dredging observations suggests that it is principally sand that has accreted in the estuary. Available evidence (Price and Kendrick, 1964) suggests a typical grain size in the estuary of 0.18 mm, which represents a fine sand material, which would be expected to behave as noncohesive material. Furthermore, a study by (Peirce et al., 1970) reports a sand fraction of 50% at Bromborough in the inner estuary, indicating noncohesive sediment transport is an important component of Mersey Estuary morphology. However, there is limited data available to calibrate precise models of non-cohesive sediment transport in Liverpool Bay. The most efficient form of examining sediment transport pathways between the estuary and wider environment is therefore to simplify or schematise the system. In the context of the current study, a schematisation of the system was achieved by assuming an inexhaustible supply of uniform sediment throughout the system. The only limiting factor on sediment transport therefore is hydrodynamic flow. Although the outputs from the modelling are more qualitative, uncertainties in the results due to parameterisation and calibration are reduced, and changes in patterns of sediment transport can be examined. 5.5. Non-cohesive bedload transport Bedload transport (qb ) of non-cohesive sediment was calculated for depth integrated flows from the 2d model using van Rijns (1984) parameterisation of full sediment transport formulae to determine a depth-averaged sediment concentration employed in bed exchange relations  2:4  1:2 U%  U% cr d50 % qb ¼ 0:005Uh ð2Þ 1=2 h ½ðs  1Þgd50 

with U% cr ¼ 0:19ðd50 Þ0:1 log10



4h d90

1785



for 100pd50 p500 mm;

ð3Þ

where U% is the depth-averaged velocity, U% cr is threshold depth-averaged current speed, h is water depth, s is the ratio of densities of grain and water (2.65), g is gravitational acceleration (9.81 m/s2), d50 is median grain diameter and d90 is 90 percentile grain size diameter. Bedload transport (qb ) of non-cohesive sediment was calculated for depth varying flows using van Rijns (1984) sediment transport formulae   0:005 d 0:2 1=2 1=2 1=2 2:4 qb ¼ 1:7 y ðy  ycr Þ ½gðs  1Þd 3 1=2 ; CD h ð4Þ where y¼

t gðps  pÞd

ð5Þ

and t ¼ pCd U 2 ;

ð6Þ

where y is the Shields parameter, ycr is threshold Shields parameter, d is grain diameter, t is bed shear stress, p is density of water (1027 kg/m3; MAFF, 1981), ps is density of sediment grains (2650 kg/m3) Cd is flow drag coefficient and U 2 is bed current velocity. The 2d and 3d potential bedload residual transport patterns are presented in the same format, as vector plots in Figs. 8 and 9, respectively. From the 2d plots it is clear that the net movement of sediment for all bathymetric configurations is seawards through the Narrows, and is in an offshore direction away from the estuary through the trained channel. Clear differences are apparent in the 3d results when compared with the 2d results, largely representing the effects of salinity-induced gravitational circulation. The net movement of sediment through the Narrows has continuously been landwards through the period 1906–1977. The net movement of sediment through the trained channel is towards the estuary at the southern end, but further away from the estuary mouth where stratification effects will have

1786

C.G. Thomas et al. / Continental Shelf Research 22 (2002) 1775–1794

Fig. 8. Vector plot of potential non-cohesive bed load sediment transport from 2d hydrodynamics results.

Fig. 9. Vector plot of potential non-cohesive bed load sediment transport from 3d hydrodynamics results.

C.G. Thomas et al. / Continental Shelf Research 22 (2002) 1775–1794

reduced the movement of sediment is away from the estuary. It is, therefore, unlikely that sediment enters the mouth of the trained channel, but the channel may act as a pathway for sediment where localised overtopping occurs nearer the estuary mouth. Although the transport patterns show some degree of variation in quantities of sediment transported, there is no significant change in net trends of the direction of sediment movement through the Narrows and trained channel for both sets of results through the period examined. The most likely source of sediment to the estuary is through Liverpool Bay, the most significant changes in potential sediment transport occurring to the west of the training walls where the strength of the ebb currents over Great Burbo Bank are weakened for 1936, enhancing flood tide transport in this area. Sediment is prevented from reaching the estuary in 1906 by a net westward movement of sediment away from the estuary mouth. By 1977 the transport of sediment through Liverpool Bay has declined significantly, and there is a net offshore movement of sediment across Liverpool Bay. Although the patterns demonstrated in Liverpool Bay are similar for 2d and 3d results due to the weakening of stratification effects in this area, it is clear that values calculated from the 2d simulation are higher than from the 3d simulation. Both sets of calculations rely on accurate estimation of bed shear stress, which is more accurately derived from 3d simulations representing flow patterns at the bed. 5.6. Non-cohesive suspended load transport In this paper bedload is primarily considered, with the underlying assumption that historical patterns of suspended load will follow similar trends as suspended load of sediment transport with a median particle size of 0.18 mm will be transported predominantly in the lower levels of flow. Suspended load transport has not been calculated for 2d model simulation results as it has been demonstrated that patterns of bedload transport differ significantly from the 3d model results and are less accurate for the purposes of calculating bed shear stress. To compare potential suspended load transport with bedload transport,

1787

the suspended load transport (qs ) of non-cohesive sediment was calculated for 3d results by taking flow results at a height of 0.4*water depth to approximate depth-averaged flow and calculating sediment transport with van Rijns (1984) parameterisation of full sediment transport formulae  2:4   U%  U% cr d50 % qs ¼ 0:012Uh ðDn Þ0:6 h ½ðs  1Þgd50 1=2 ð7Þ with  Dn ¼

gðs  1Þ n2

1=3 d;

where U% is the depth-averaged velocity, U% cr is threshold depth-averaged current speed, h is water depth, s is the ratio of densities of grain and water (2.65), g is gravitational acceleration (9.81 m/s2), d50 is the median grain diameter and Dn is dimensionless grain size. It is evident that patterns of suspended sediment transport shown in Fig. 10 exhibits similar characteristics to bedload transport (see Fig. 9), although the quantities transported are significantly greater. The only clear pathway for sediment to enter the estuary is through Liverpool Bay in 1936 according to 3d modelling results. This indicates that salinity-induced gravitational circulation has a significant effect on net sediment transport at a height of 0.4* the depth in addition to bed layer effects. The potential transport of non-cohesive material as suspended load into the estuary enables significantly greater quantities of sediment to enter the estuary than through bedload transport alone, and combined suspended and bedload transport rates are required to account for the magnitude of observed accretion in the estuary. Rates of sediment transport calculated are potential rates and rely on the assumption that there is an unlimited supply of uniform material present in Liverpool Bay for transportation according to hydrodynamic conditions. In reality this is unlikely to be the case, as there are spatial variations in sediment grain diameter and availability of sediment. Although it is evident that suspended sediment transport has potentially had

C.G. Thomas et al. / Continental Shelf Research 22 (2002) 1775–1794

1788

Fig. 10. Vector plot of potential non-cohesive suspended load sediment transport from 3d hydrodynamics results.

a greater effect on morphological evolution of the estuary system, the calculations of suspended sediment transport are a significant oversimplification of suspended sediment transport processes due to the Lagrangian nature of suspended load.

6. Discussion 6.1. The nature of morphological change The Mersey Estuary has exhibited a largely consistent, continuous trend of accretion between 1906 and 1977 followed by a relatively small increase in estuary capacity indicative of erosion between 1977 and 1997. Derivation of a sediment budget accounting for the effects of dredging and reclamation activity in the estuary in Table 1 indicates with greater clarity morphological trends in the estuary. The annual flux of sediment into the estuary is consistently >1.5 Mm3 between 1906 and 1977. Between 1977 and 1997, however, annual sediment fluxes declined dramatically to 0.2 Mm3 indicating that increase in estuary volume

through this period may be largely attributed to dredging activity. Thus, it appears the estuary has attained a near stable state regarding net sediment flux, although localised erosion and accretion within the estuary may still occur. Although the quantities of dredged material deposited within the estuary were not available for flux calculations for years prior to 1956 they are unlikely to be >25% of material dredged so annual sediment fluxes into the estuary would remain significantly higher than for 1977–1997. 6.2. The causes of morphological change The sustained nature of morphological change between 1906 and 1977 suggests that perturbation and subsequent recovery associated with anthropogenic activity at the end of the 19th and beginning of the 20th century has had the single most important influence on the estuary over the last 100 years. Although morphological evolution in the Mersey has been linked to the construction of training walls along the navigation channel, the onset of morphological change cannot be defini-

C.G. Thomas et al. / Continental Shelf Research 22 (2002) 1775–1794

1789

Fig. 11. Bathymetric changes in Liverpool Bay.

tively attributed to training wall construction. The construction of the training walls was itself a response to changes in sediment transport patterns in Liverpool Bay that may be related to the onset of morphological change in the Mersey Estuary, since it was constructed to stabilise the position of the low water channel. Earlier anthropogenic activity represents a possible cause of morphological perturbation in Liverpool Bay prior to training wall construction, particularly dredging of the bar at the seaward end of the navigation channel in 1890 from a depth of 4 m below Low Water Springs to 10 m below Low Water Springs (Cashin, 1949). The probable impact of dredging in the navigation channel was to increase the ebb tidal flow over the bar resulting in gyres to the north and south of the seaward entrance to the channel (HR Wallingford, 2000). Changes in flow at the mouth of the navigation channel is associated with the formation of sandbanks,

Taylors Spit and Askew Spit, at the entrance to the navigation channel evident on historical charts. Following construction of the training walls, however, it is evident that there have been significant changes in sediment transport patterns in Liverpool Bay. The changes are related to the effect of the training walls constraining ebb flow from the estuary and concentrating flow of water exiting the estuary within the trained low water channel. Ebb flow in the Rock Channel, and flow over Burbo Bank was thus reduced and the subsidiary channels became more flood dominant as reflected in changes in sediment transport patterns presented in Figs. 9 and 10. As a result sediment accreted in these channels as demonstrated in bathymetric changes presented in Fig. 11. The resulting expansion and increase in height of Great Burbo Bank caused sediment to overtop the training wall locally and also increased

C.G. Thomas et al. / Continental Shelf Research 22 (2002) 1775–1794

1790

supply of sediment to the mouth of the estuary. Density currents in the Narrows extending into a trained section of the Crosby Channel then transported material into the estuary. 6.3. Sediment source The source of sediment entering the Mersey Estuary represents a key issue for developing conceptual understanding of the morphological functioning of the system. The scale of sedimentation experienced in the Mersey Estuary indicates it is most probable sediment entered the estuary system from seaward sources via the salinityinduced gravitational circulation in the Narrows. Potential annual sediment flux at the estuary mouth calculated from 3d results indicate a potential net annual flux of sediment into the estuary for 1906, 1936 and 1977 demonstrating the potential for the estuary to import sediment. Making the assumption that sediment is only transported on spring tides an approximate annual sediment flux into the estuary has been calculated in Table 4. Calculations are based upon the integration of sediment transport calculated from 3d simulation, adding computed bed and suspended load transport moving landwards across cross-section 1 shown in Fig. 12, and multiplying by 365 to represent the annual number of spring tides. The values are approximate as sediment transport calculations have significant error bounds, van Rijn (1993) found for example that total sediment transport calculated according to the parameterisation employed in this study was

Table 4 Non-cohesive sediment flux through the mouth of the Mersey Estuary (Section 1) calculated from 3d simulation results Year

1906 1936 1977 a

Net potential spring tide residual bedload flux through section (m3) Bed load

Suspended load

Total flux

175 191 125

4680 5103 3356

4855 5294 3481

Assumes 365 spring tides per year.

Annual flux (Mm3)a

1.772 1.932 1.271

Fig. 12. Location of cross sections in Liverpool Bay for potential sediment flux calculations.

accurate to a factor of 2 for 76% of cases studied. Nevertheless, the values calculated for annual flux of sediment through the estuary mouth in a landward direction match reasonably well with the annual sediment flux derived from the sediment budget in Table 1. The results indicate bed and suspended load transport of non-cohesive sediment under tidal forcing accounts for a substantial proportion of sediment transported into the estuary. The evidence presented in this study suggests that the most probable source of sediment causing large-scale accretion in the estuary is the east side of Liverpool Bay. Examination of potential sediment transport patterns in Liverpool Bay indicates hydrodynamic conditions have altered changing potential for sediment to be transported to the estuary mouth. Following changes to hydrodynamic conditions and bathymetry in Liverpool Bay, there is a clear eastward movement of sediment through Liverpool Bay to the mouth of the Mersey Estuary. Examination of changes in the bathymetry of Liverpool Bay (see Fig. 11) indicate significant erosion of Great Burbo Bank between 1906 and 1977, providing a substantial source of sediment for transportation into the estuary. Furthermore, plots of sediment transport vectors based upon hydrodynamic model results (Figs. 9 and 10) demonstrate little potential for

C.G. Thomas et al. / Continental Shelf Research 22 (2002) 1775–1794 Table 5 Non-cohesive sediment flux across Liverpool Bay (Section 2) calculated from 3d simulation results Year

1906 1936 1977 a

Net potential spring tide residual bedload flux through section (m3) Bed load

Suspended load

Total flux

34 78 8

592 1264 190

626 1342 198

Annual flux (Mm3)a

0.228 0.490 0.072

Assumes 365 spring tides per year.

sediment transport along the Formby coastline although the model does not accurately represent longshore drift, which requires an approach simulating refraction of tidal and wave currents from the shoreline. Net annual flux calculations of sediment crossing Section 2 in Fig. 12 in the direction of the estuary based upon integration of sediment transport calculated from 3d simulations of bed and suspended load transport across the section multiplied by 365 are presented in Table 5. It is evident that sediment flux across this section has varied considerably between 1906 and 1977. 1936 exhibits a significantly larger annual sediment flux than 1906 and 1977 indicating that transport of sediment upon the flood tide was significantly greater in 1936, most probably due to a reduction in ebb flow across the eastern side of Liverpool Bay as ebb tidal flow became constrained within the trained channel. By 1977 it appears that the bathymetry of Liverpool Bay had adjusted to induce a reduction of flood tidal velocities due to the erosion of Great Burbo Bank evident in Fig. 11, reducing flood tidal velocities. The calculated annual landward sediment flux across Section 2 is not sufficient to account for the net annual sediment flux calculated in the sediment budget in Table 1. However, it is likely that the area of Liverpool Bay around Section 2 is exposed to wave activity which could have significantly enhanced sediment transport due to wave stirring effects increasing the concentrations of sediment in suspension which could be transported by tidal currents.

1791

6.4. Estuary response to perturbation It has been suggested in literature that estuaries and inlets can respond to perturbation by adjusting system geometry to alter non-linear tidal behaviour to achieve an equilibrium state. Changes in integrated geometrical parameters in the Mersey Estuary indicate a morphological response in the Mersey Estuary to perturbation consistent with adaptation of tidal asymmetry characteristics to prevent sediment import into the estuary. Analysis of Dronkers g parameter indicated a decline through the period 1906–1997 to a value closer to 1 representing a decrease in flood dominance towards an approximate balance between flood and ebb tidal flow. In addition, changes in flow characteristics in the Narrows derived from computational hydrodynamic results indicate a reduced potential for landward sediment transport through the Narrows (Figs. 7 and 8). However, the Narrows are largely inerodible and there is little adjustment of system geometry in this area. Changes in tidal flow characteristics within the Narrows represented in 2d and 3d computations of hydrodynamics most probably result from changes in boundary forcing caused by bathymetric changes in Liverpool Bay distorting the progression of the shallow water tidal wave into the estuary. Despite changes in hydrodynamic conditions in the Narrows, calculation of net non-cohesive sediment fluxes indicate the estuary has exhibited continuous potential to import sediment. Thus, even if sediment regime landwards of the Narrows, where system geometry has greater ability to adapt and alter hydrodynamic regime, adjusted to export sediment, sediment imported through the Narrows would still be deposited within the estuary where the Narrows currents slow or transported up the flood channel of the Middle Deep. The geological constriction of the Narrows and the existence of gravitational circulation as a mechanism for importing sediment has remained relatively unaltered by anthropogenic activity and morphological change, preventing adaptation of sediment regime within the estuary to achieve a new equilibrium state. The estuary instead appears to have attained a new equilibrium state due to a restriction of

1792

C.G. Thomas et al. / Continental Shelf Research 22 (2002) 1775–1794

sediment supplied to the estuary mouth. Thus, feedback between hydrodynamic conditions and bathymetric configuration in Liverpool Bay controlling supply of sediment to the estuary has acted as a dominant control upon morphological change and has been responsible for causing and ending accretion.

*

7. Conclusions Significant benefits are realised from analysing the nature and causes of morphological change in an estuary using a combination of approaches including analysis of bathymetric data, calculation of analytical parameters and hydrodynamic simulations. This paper documents the nature of longterm change in the Mersey Estuary and develops a conceptual understanding of the causes of morphological evolution. The major findings of this investigation include the following: *

*

The Mersey Estuary has experienced substantial morphological change over the period 1871– 1997, with significant accretion occurring between 1906 and 1977. Derivation of a sediment budget of the estuary substantially enhanced interpretation of bathymetric data indicating that the estuary attained a state of stability between 1977 and 1997 in terms of a net sediment flux through the estuary mouth of approximately zero. Estuary volume increased between 1977 and 1997 as a result of dredging activity within the estuary, which is indicative of localised erosion and accretion, so the estuary cannot be regarded as completely stable. Estuarine response to perturbation induced by anthropogenic activity in Liverpool Bay occurred over a period of E70 years. The duration of response meant several anthropogenic impacts were present within the estuary system at the same time, with an identifiable order of impact dominance. The response to perturbation was the dominant impact between 1906 and 1977 and was superimposed upon the effects of other anthropogenic activity such as dredging in the estuary. Although areas within the estuary have been dredged continuously

*

*

*

between 1906 and 1997, dredging activity has only had a dominant impact upon the net behaviour of the system as net sediment flux into the estuary has declined. The cause of sedimentation in the estuary through the period 1906–1977 has been related to construction of training walls in Liverpool Bay causing changes in sediment transport patterns outside the estuary which increased supply of sediment to the estuary mouth. Liverpool Bay has, however, experienced other anthropogenic activity linked to changes in hydrodynamic regime in Liverpool Bay such as dredging of the bar at the end of Queens Channel and dredging of Crosby Channel. The precise impact of training wall construction cannot be differentiated from these impacts, and changes in the Mersey Estuary cannot be definitively attributed to training wall construction; it is most probably the result of cumulative impacts of anthropogenic activity in Liverpool Bay. Differences between 2d and 3d results demonstrated the importance of salinity-induced gravitational circulation in the Narrows and provide the only identified means of transporting sediment into the estuary. 3d representation of the estuary bed layer also enabled more accurate estimation of bed shear stress accounting for overestimation of sediment flux calculations based upon 2d simulations compared with calculations based upon 3d simulations. Changes in flow characteristics within the Narrows indicated a reduction in potential capacity to import sediment between 1906 and 1977. Due to the inerodible nature of the Narrows changes in flow characteristics are most probably induced by changes in tidal forcing resulting from bathymetric change in Liverpool Bay. Overall, however, identified changes in flow characteristics have had little impact upon the potential import of sediment through the Narrows according to sediment flux computations. In the inner estuary which can adjust more readily to changes in forcing processes, changes in geometric parameters were reconcilable with a morphological response to perturbation

C.G. Thomas et al. / Continental Shelf Research 22 (2002) 1775–1794

*

indicated by Dronkers (1998) parameter. This supports the theory that estuaries respond to perturbation by altering geometric parameters to adjust non-linear tidal processes in order to attain a stable state. However, in the case of the Mersey sediment transported into the estuary has been dominated by the influence of the Narrows which continuously exhibited potential to import sediment. The adjustment of estuarine system geometry has therefore not been a dominant control upon net morphological behaviour. Changes in movement of sediment across Liverpool Bay in response to changes in hydrodynamic regime were found to be the principal cause and controlling influence upon accretion in the Mersey Estuary. The effects of training wall construction and dredging of the Crosby Channel were to increase ebb tidal flow through the Crosby Channel and reduce ebb tidal flow over the Great Burbo Bank. As a result flood tidal flow over Great Burbo Bank became increasingly dominant enhancing the movement of non-cohesive sediment towards the estuary mouth. Stability was attained as a result of bathymetric adjustment in Liverpool Bay changing hydrodynamic flow patterns and reducing the dominance of flood tidal noncohesive sediment transport over Great Burbo Bank, reducing the supply of sediment to the estuary mouth.

Acknowledgements This research conducted for this paper was based on the provision of historical hydrographic charts of the Mersey Estuary recorded by the Mersey Dock and Harbour Company, used with kind permission of the Acting Conservator of the River Mersey, Mr. Fraser Clift.

References Cashin, J.A., 1949. Engineering works for the improvement of the Mersey. Journal of the Institute of Civil Engineers 32, 296–355.

1793

Castaing, P., Allen, G.P., 1981. Mechanisms controlling seaward escape of suspended sediment from the Gironde: a macrotidal estuary in France. Marine Geology 40, 101–118. Draper, L., Blakey, A., 1969. Waves at the Mersey Bar Light Vessel. (National Institute of Oceanography Report A37) National Institute of Oceanography, Wormley. Dronkers, J., 1998. The morphodynamics of the Dutch Delta. In: Dronkers, J., Scheffers, M.B.A.M. (Eds.), The Physics of Estuaries and Coastal Seas. Balkema, Rotterdam, pp. 297–304. Dyer, K., 1997. A Physical Introduction to Estuaries. Wiley, London. Fitzgerald, D.M., Nummedal, D., Kana, T.W., 1976. Sand circulation at Price Inlet, SC. In: Proceedings Coastal Engineering Conference, Honolulu, pp. 1868–1880. Friedrichs, C.T., Aubrey, D.G., 1988. Nonlinear tidal distortion in shallow well mixed estuaries: a synthesis. Estuarine Coastal and Shelf Science 27, 521–545. Heaps, N.S., Jones, J.E., 1977. Density currents in the Irish Sea. Geophysical Journal of the Royal Astronomical Society 1951, 393–429. Hervouet, J.M., Van Haren, L., 1994. Telemac 2-D Version 3.0 Principle Note, EDF-DER Report HE-43/94/052/B. HR Wallingford, 2000. Two and three dimensional modelling of sediment transport mechanisms in the Mersey Estuary. HR Wallingford Report TR 109, HR Wallingford, UK. MAFF, 1981. Atlas of the seas around the British Isles. Ministry of Agriculture Fisheries and Food, Lowestoft, UK. HMSO. Peirce, T.J., Jarman, R.T., de Turville, C.M., 1970. An experimental study of Silt Scouring. Proceedings of the Institute of Civil Engineers 49, 231–243. Price, W.A., Kendrick, M.P., 1963. Field and model investigation into the reasons for siltation in the Mersey Estuary. Proceedings of the Institute of Civil Engineers 24, 473–517. Price, W.A., Kendrick, M.P., 1964. Discussion paper no. 6639 on ‘‘Field and model investigation into the reasons for siltation in the Mersey Estuary’’. Proceedings of the Institute of Civil Engineers 27, 613–647. Ramster, J.W., 1975. Salinity, temperatures and density distributions in Liverpool Bay. DoE report ‘‘Out of Sight Out of Mind’’. HMSO, UK. Sly, P.G., 1966. Marine geological studies in the Irish Sea and adjacent estuaries with special reference to sedimentation in Liverpool Bay and the River Mersey. Unpublished Ph.D. Thesis, University of Liverpool. Van Dongeren, A.R., De Vriend, H.J., 1994. A model of morphological behaviour of tidal basins. Coastal Engineering 22, 287–310. van Rijn, L.C., 1984. Sediment transport. Part I: bedload transport. Part II: suspended load transport. Part III: bed forms and alluvial roughness. Proceedings ASCE Journal of Hydraulics Division 110 (HY10) 1431, (HY11) 1613–1641, (HY12), 1733–1754.

1794

C.G. Thomas et al. / Continental Shelf Research 22 (2002) 1775–1794

van Rijn, L.C., 1993. Principles of Sediment Transport in Rivers, Estuaries and Coastal Seas. Aqua Publications, Amsterdam. Water Pollution Research Laboratory, 1938. The Effect of Discharge of Crude Sewage into the River Mersey on the Amount and Hardness of the Deposit in the Estuary. Water Pollution Research Laboratory Technical Paper no. 7, H.M.S.O., London. West, J.R., 1980. Mersey Estuary water level measurements July 1980. Department of Civil Engineering University of Birmingham Report, Unpublished.

Winters, A., 1984. Suspended sediment concentrations in the Mersey Estuary. HR Wallingford Report DE 70, HR Wallingford, UK. Woodworth, P.L., 1999. High waters at Liverpool since 1768: the UK’s longest sea level record. Geophysical Research Letters 26 (11), 1589–1592. Wright, J.E., Hull, J.H., McQuillin, R., Arnold, S.E., 1971. Irish Sea Investigations 1969–1971. Institute of Geological Sciences Report, 71/19.