Estuarine, Coastal and Shelf Science 94 (2011) 343e354

Contents lists available at ScienceDirect

Estuarine, Coastal and Shelf Science journal homepage: www.elsevier.com/locate/ecss

Contribution of hydrodynamic conditions during shallow water stages to the sediment balance on a tidal flat: Mont-Saint-Michel Bay, Normandy, France R. Desguée a,1, N. Robin a, 2, L. Gluard a, O. Monfort a, *, E.J. Anthony b, F. Levoy a a

Unité Morphodynamique Continentale et Côtière, Université de Caen - Basse-Normandie, 2-4 rue des Tilleuls, 14000 Caen Cedex, France Centre Européen de Recherche et d’Enseignement des Géosciences de l’Environnement, Université d’Aix-Marseille, Institut Universitaire de France, Europôle Méditerranéen de l’Arbois, 13545 Aix-en-Provence cedex 04, France b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 15 July 2011 Accepted 17 July 2011 Available online 28 July 2011

Field measurements were conducted in Mont-Saint-Michel Bay, a megatidal embayment (spring tidal range of 15 m), in order to monitor, over the course of a tidal cycle, sediment transport variability due to waves and tides on the upper part of a tidal flat characterised by shallow water depths. Sensors used to measure currents, water depth and turbidity were installed just above the bed (0.04 m). Two experiments were conducted under contrasting hydrodynamic conditions. The results highlight wave activity over the tidal flat even though observed wind waves were largely dissipated due to the very shallow water depths. Very high suspended sediment concentrations (up to 6 kg/m3) were recorded in the presence of wave activity at the beginning of the local flood, when significant sediment transport occurred, up to 7 times as much as under conditions of no wave activity. This influence may be attributed to the direct action of waves on bed sediments, to wave-induced liquefaction, and to the erosive action of waves on tidal channel banks. The sediment composition, comprising a clay fraction of 2e5%, may also enhance sediment transport by reducing critical shear stress through the sand lubrication effect. The results also show that antecedent meteorological conditions play an important role in suspended sediment transport on the tidal flat. Total sediment flux directions show a net transport towards the inner part of the bay that contributes to deposition over the adjacent salt marshes, and this tendency also prevails during strong wave conditions. Such sediment transport is characterised by significant variability over the course of the tidal cycle. During fair and moderate weather conditions, 83% and 71% of the total flux was observed, respectively, over only 11% and 28% of the duration of the local tidal cycle and with water depths between 0.04 and 0.3 m. These results suggest that in order to improve our understanding of sediment budgets in this type of coastal environment, it is essential to record data just at the beginning and at the end of tidal submergence close to the bed. Ó 2011 Elsevier Ltd. All rights reserved.

Keywords: tidal flat megatidal environment sediment transport turbidity wave action shallow water France Mont-Saint-Michel Bay

1. Introduction Due to their position at the interface between marine and terrestrial environments, tidal flats are highly dynamic environments. Sediment movement on tidal flats and the consecutive bed changes have strong societal and economic impacts because of * Corresponding author. Present address: CREC, station marine de Luc-sur-Mer, 54 rue du Docteur Charcot, BP 49, 14530 Luc-sur-Mer, France. E-mail addresses: [email protected] (R. Desguée), nicolas.robin@ univ-perp.fr (N. Robin), [email protected] (L. Gluard), olivier.monfort@ unicaen.fr (O. Monfort), [email protected] (E.J. Anthony), [email protected] (F. Levoy). 1 Present address: Syndicat Mixte pour le Rétablissement du Caractère Maritime du Mont-Saint-Michel, 2 rue du Prieuré, 50170 Ardevon, France. 2 Present address: Laboratoire CEFREM UMR 5110, Université de Perpignan Bat U, 52 Avenue Paul Alduy, 66860 Perpignan Cedex, France. 0272-7714/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.ecss.2011.07.010

their commercial (oyster and mussel cultures), recreational, and ecological importance (Kuriyama et al., 2005). Sandy tidal flats are complex environments and studies are often highly specific in character, due to the difficulty of investigating simultaneously the broad range of parameters affecting these flats. These include potentially strong wind wave activity in what are generally lowenergy environments, variable tidal flows, drainage, a marked degree of sediment heterogeneity involving a pseudo-cohesive mixture of sand, silt and clay, and a high degree of bioturbation (Malvarez et al., 2004; Anthony, 2009). Tidal flats are commonly found in macro-tidal environments, with a tidal range higher than 4 m, and several examples have been described from sites with exceptional tidal ranges such as the Bay of Fundy in Canada (18.5 m) (Davidson-Arnott et al., 2002), the Severn estuary in England (16.5 m) (Kirby and Kirby, 2008), and MontSaint-Michel Bay in France (15 m) (Le Rhun, 1982; Tessier, 1990;

344

R. Desguée et al. / Estuarine, Coastal and Shelf Science 94 (2011) 343e354

Larsonneur, 1994; Bonnot-Courtois et al., 2002; Cayocca et al., 2006; Billeaud et al., 2007). These environments are commonly characterised by semi-diurnal tides. Maximum current speeds, generally occurring at about the mid-tide level, commonly exceed 1 m/s and can reach 2 or 3 m/s during very large spring tides, especially in tidal channels. Speeds, however, may be close to zero during slack water (Bassoulet et al., 2000; Le Hir et al., 2000; Lee et al., 2004; Yang et al., 2007). In these high tide-range environments, the influence of waves is generally considered as less important than in microtidal environments, mainly because of the strong tides and the sheltered or commonly highly embayed setting of tidal flats (Eisma, 1998), with wide, low-sloping and ultra-dissipative intertidal zones. However, although tidal currents play a greater role both in absolute terms and relative to local waves, the role of waves can, nevertheless, be important in the intertidal zone (Reineck, 1967; Amos, 1995; Lee et al., 2004; Paphitis and Collins, 2005; Anthony, 2009). JanssenStelder (2000) and Kim (2003) have suggested that the development of tidal flats is mainly controlled by the interaction between tidal currents and wind waves, the latter dominating the lower seaward front of the flats where erosion results in a concaveeupward profile, while the higher landward part is tidally dominated with a convexeupward profile that is believed to have a tendency to accrete over time (Kirby, 2000; Le Hir et al., 2000; Roberts et al., 2000; Pritchard et al., 2002). More recently, Lambrechts et al. (2010) have shown that wave-induced liquefaction of the fine sediment bed is the main process inducing erosion, and this indeed highlights the potentially strong influence of waves even in settings where overall wave activity is low. Sedimentation in the upper part of the intertidal zone depends largely on settling and scour-lag asymmetries (Talke and Stace, 2008). Sediment transported onto the tidal flat during flood tides can settle and be deposited around the high water slack. Sediments deposited on the bed may partially consolidate during slack water such that equivalent ebb velocities may not be sufficient to resuspend them (Dyer et al., 2000; Lawrence et al., 2004; Lee et al., 2004; Ralston and Stacey, 2007; Anthony et al., 2008). Under calm weather conditions, minor sediment transport can usually be

observed (Ralston and Stacey, 2007). Small waves can erode large amounts of sediments, which are then transported by tidal currents (Anderson and Black, 1981). However, stormy conditions with strong winds can generate large wind waves capable of removing large quantities of sediments (Lee et al., 2004; Paphitis and Collins, 2005; Yang et al., 2007), even when the wave height and induced bottom shear stress are closely limited by the water depth (Le Hir et al., 2000). Changes in hydrodynamic conditions during both neap-spring tide cycles and storm activity have also been shown to affect the balance of erosion and sedimentation processes (Ryan and Cooper, 1998; Ridderinkhof et al., 2000) in the upper intertidal zone. The literature on sand flats shows a lack of studies on sedimentary processes carried out over an entire semi-diurnal tidal cycle and involving measurements very close to the bed, where suspension from the bed must dominate. Such measurements are important to estimate bed shear stress as mentioned by Andersen et al. (2007) for a muddy microtidal environment and in the establishment of sediment balances prevailing over tidal flats (Lambrechts et al., 2010). The aim of this paper is to investigate sediment transport on a tidal flat composed of fine sand and located in a megatidal setting associated with strong tidal currents and episodic strong wave conditions. Hydrodynamic and suspended sediment data obtained from measurements conducted in very shallow waters, at water depths below 0.04 m, have enabled quantification of sediment fluxes over a semi-diurnal spring tidal cycle during both fair and moderately stormy conditions. Flux directions and intensities were also calculated to appreciate their temporal variability and their contributions to the sediment balance. 2. Study area The study was conducted in Mont-Saint-Michel Bay, worldfamous for its monastery, located in the English Channel between the Cotentin Peninsula (oriented NeS) and the Brittany coast (oriented EeW) (Fig. 1) in France. Mont-Saint-Michel Bay is characterised by a large (w500 km2) sandy tidal flat cut by three small rivers (the Sée, the Sélune and the Couesnon). The sandy tidal flat

Fig. 1. Study area. Photograph shows the instrument deployment site (labelled A) close to the world-famous Mont-Saint-Michel Monastery.

R. Desguée et al. / Estuarine, Coastal and Shelf Science 94 (2011) 343e354

occupies the lower and mid-intertidal zones, and is backed by high tide salt marshes dissected by numerous tidal channels. The salt marshes front polders built over the two last centuries. The low and mid-intertidal zones are characterised by fine sand with a mean grain size of about 150 mm. Current ripples with wavelengths and amplitudes ranging from 1 to 10 cm are common in this zone. The abundant fine sand and the relatively energetic tidal conditions prevailing in this zone preclude the development of hydraulic dunes. However, mud is sometimes deposited in the troughs of the ripples or in wider topographic depressions (Bonnot-Courtois et al., 2002). Channel bank surfaces are generally characterised by a smooth, flat, upper flow-regime surface. The upper intertidal area consists of fine minerogenic sediment (D50: 110 mm) that is locally referred to as “tangue”. This dominantly very fine sand comprises between 2 and 5% of clay, and the overall mud content (silt and clay) in the study area ranges from 20 to 25%. Following the sand-silt-clay triangle of Van Ledden et al. (2004), “tangue” is a non-cohesive sand-dominated sediment (textural bed type I). Macrozoobenthos is rare on this part of the intertidal area and consists mainly of Corophium Arenarium and Macoma balthica (Thorin et al., 2001). The channels dissecting the upper intertidal zone and salt marshes form sheltered low-energy environments susceptible to rapid infill and accretion. Salt marsh vegetation consists of Spartina Anglica and Puccinellia Maritima which favour accretion and the storage of suspended sediments when covered by tides (Neumeier and Amos, 2006). The regional setting of the study site is a fine example of a megatidal environment (Larsonneur, 1994; Levoy et al., 2000; Robin et al., 2009). The offshore area is characterised by very complex hydrodynamic conditions. The tidal wave propagating eastward from the Atlantic Ocean is reflected by the Cotentin Peninsula, generating a standing tidal wave especially within Mont-Saint-Michel Bay where the tidal range is about 12 m during mean spring tides and attains 15 m during very large spring tides (Le Rhun, 1982). The tidal wave is dominated by the M2 (semidiurnal) harmonic, the amphidromic point of which is a virtual one located on land in southwest England (Pingree and Griffiths, 1979). The bay tidal prism landward of a line running from the Pointe du Grouin, through the Chausey Islands to Granville (Fig. 1), is about 5.109 m3 at each spring tide. During the flood, strong tidal currents progress towards the inner bay with speeds exceeding 1 m/s close to the Pointe du Grouin. Near Mont-Saint-Michel Monastery, current speeds usually attain between 0.3 and 0.7 m/s (Laboratoire Central d’Hydraulique de France, 1977). In the river channels, they can even reach 2.2 m/s at the beginning of the rising tide for spring conditions. Currents during the flood limb are always stronger than those of the ebb (Bonnot-Courtois et al., 2002), thus reflecting flood-dominated asymmetry. The duration of the ebb is 1.17 times longer than that of the flood (Migniot, 1997). Prevailing winds are from west to northwest. During storms, strong winds may also rarely come from southwest to south, generating wind waves. Wave propagation is complicated by the shoreface bathymetry, by the Channel Islands, and by the numerous shoals and islets, which result in a decrease in wave heights over the shoreface (Levoy et al., 2000). As a result, low wave-energy conditions prevail near Mont-Saint-Michel Monastery. Recorded wave heights near the study area between October 2006 and January 2009 are less than 0.1 m 71.6% of the time and between 0.1 and 0.2 m 17.6% of the time. Wave heights larger than 0.4 m were observed only 2.3% of the time. Waves come mainly from a west to northeast window (270
e45
) (78% of the observations). Peak periods (Tp) range from 2.5 to 17 s, reflecting a mix of wind waves (Tp between 2.5 s and 6 s 46% of the time) and swell.

345

3. Methods 3.1. Instrument deployment The study area is situated east of Mont-Saint-Michel Monastery (Fig. 1), where a salt marsh has begun to develop. Several short experiments, each conducted over one semi-diurnal tidal cycle, were carried out from 7 December 2006 to 23 November 2007. The experimental site was located on the upper part of the sand flat at the limits of pioneer vegetation, corresponding to the mean high tide level. The methodology was identical for each experiment. 3.2. Collection and analysis of hydrodynamic data Hydrodynamic measurements were carried out using a microAcoustic Doppler Velocimeter (2D micro-ADV Sonteck). The sensor was installed in a way as to measure currents 0.04 m above the bed, and the data were collected continuously at 4 Hz. Velocity measurements from the ADV tended to become noisy in highly turbulent flows. Signal correlation values recorded by the ADV were used to identify such potentially inaccurate data. When signal correlation for a given acoustic beam was less than 85%, the data were discarded. Mean current velocities were computed as 1-min averages. On the same frame, a pressure sensor was mounted 0.04 m above the bed. Data were collected in continuous mode at a 4 Hz frequency. Wave characteristics were processed by standard spectral analysis using Fast Fourier Transforms. The Fournier coefficients of the free surface elevation fluctuations were obtained from the corresponding ones computed from the pressure time series. To this end, the frequency-dependent transfer function inferred from linear wave theory was used. In order to identify the roles of waves and currents in sediment transport, the current- and wave-induced bed shear stresses were calculated for each experiment. Critical bed shear stress was determined by Migniot (1997) from flume experiments for the local “tangue” sediment without taking into account potential biological mediation, but with a good representation of the sediment sizes measured in the field. Due to its complex composition, the critical bed shear stress of “tangue” is close to 1 N/m2 (Migniot, 1997). Le Hir et al. (2008) used an erodimeter to calculate critical shear stresses of sediments in the western part of Mont-Saint-Michel Bay. The erosion threshold for natural sand/mud mixtures varied from 0.25 N/m2 to 1.5 N/m2 with clay fractions between 2.5% and 5%, which is essentially the case in the field site near the Mont-SaintMichel Monastery. These authors found no correlation between the erosion threshold and the concentration (for concentrations between 300 and 950 kg/m3). Currents and wave-induced bed shear stresses were computed, following the Van Rijn (1993) model, using:





!2

sc ¼ r$fc $ V r þ ! ur

2

. .
8 and sw ¼ r$fw $Ud2 4

! where V r , the depth-averaged velocity vector in the main current direction, is obtained by numerical integration of a logarithmic velocity distribution over the depth inferred from the measured ! velocity and computed bed roughness heights (Van Rijn, 1993), u r is the time-averaged and depth-averaged return velocity due to wave action, Ud is the near-bed peak orbital velocity, and fc and fw friction coefficients due to currents and waves. 3.3. Collection and analysis of suspended sediment One Optical Backscatter Sensor (OBS) was mounted 0.04 m above the bed on a rig and also programmed for continuous

346

R. Desguée et al. / Estuarine, Coastal and Shelf Science 94 (2011) 343e354

recordings at 4 Hz. No significant scour signs were observed around the stem during inspections at low water. OBS is designed for measuring turbidity and suspended solid concentrations by detecting infrared radiation scattered from suspended matter at angles between 140
and 160
. Optical backscatter measurements are very sensitive to particle size (Bunt et al., 1999), and the shape and the composition of the particle also play an important role (Green and Boon, 1993). Indeed, OBS are more sensitive to mud than to sand because of the larger surface area to mass ratios for mud particles than for sands. Consequently, OBSs need to be calibrated for each experiment using suspended sediments from the field site. During laboratory experiments, the average OBS response was related to the known suspended concentrations of sediments taken from the study area. Calibration coefficients were obtained both from linear functions and 2nd order polynomials. Although the OBS response is theoretically linear, second order polynomial functions were used to obtain the reported suspended sediment concentrations because they were significantly better. 4. Results The two experiments reported here were undertaken at the same site (Fig. 2, point A) located on the upper part of the sand flat, at the limits of the pioneer vegetation, during a mean spring tidal cycle, respectively on 14 May 2007 (experiment 1) and 22 November 2007 (experiment 2). Tidal conditions for the two experiments were similar (tidal ranges close to 9.4 m), but weather and wave conditions differed considerably. 4.1. General weather and hydrodynamic conditions During experiment 1, the wind blew at 10 m/s from w280e315
. During the five days prior to the experiment (Fig. 3), southwest winds prevailed and wind speeds varied from 3 to 13 m/s. Maximum wave heights were observed at high tide (0.27 m) and decreased with the water level (Fig. 4a). Peak periods were from 2.9 to 5.6 s and waves came mainly from 280
. Although significant wave heights in the study area are weak, values up to 0.2 m are rarely observed in this part of the bay (11% of the time) and can therefore be defined as typical of moderate to stormy conditions for the study area. Mean current velocities were below 0.2 m/s. Maximum mean velocities were observed at the beginning and at the end of the tidal cycle in shallow water (0.2 m/s in a water depth, d, of 0.6 m during the flood; 0.17 m/s in a water depth of 0.06 m during the ebb) (Fig. 4b). An asymmetric-shaped velocity profile was observed with flood intensities higher than those of the ebb. On the whole, the area was submerged for 122 min. During the five days preceding experiment 2, wind speeds decreased sharply from 10.2 m/s to a minimal value of 3 m/s at the end of the experiment, with mean values close to 4.8 m/s (Fig. 5). Offshore winds blew almost exclusively from the south (w170
).

elevation (m)

P 16 14 12 10 8 6 4 2 0

SW

SM

SMF A

Consequently, significant wave heights were always below 0.05 m in spite of a north wind rotation on 22 November (from w345
), but with weaker intensities. This experiment was, thus, characterised by fair-weather conditions. Mean current velocities were at a minimum at high tide (0.05 m/s) and attained a maximum at the beginning and at the end of the tidal cycle, respectively 0.46 m/s and 0.23 m/s with water depths of 0.15 m and 0.06 m (Fig. 6). The total duration of submersion of the area was about 100 min. The shape of the current curve presented here is representative of those obtained in the course of several other experiments during fairweather conditions. 4.2. Suspended sediment concentrations (SSC) During experiment 1 (moderate to stormy conditions), SSC values were large (Fig. 4b). A strong peak reaching 6 kg/m3 prevailed for water depths smaller than 0.4 m (first 15 min). SSC values then fell below 3 kg/m3 throughout the rest of the tidal cycle. In very shallow water depths, below 0.07 m just at the end of the ebb tide, the SSC values were weak, below 2 kg/m3. No peak was observed at the end of the tidal flooding. During experiment 2 (fair-weather conditions), the SSC curve presents the same pattern as that of experiment 1 but with weaker intensities (Fig. 6b). A small peak was observed at the beginning of the flood (2.5 kg/m3 during 6 min), followed by a decrease down to 0.5 kg/m3. This value then remained constant during 90 min, until the end of the tidal cycle. Strong differences exist between these two experiments. Although the SSC curves exhibit a similar pattern, SSC were 2.4 times larger during experiment 1 for the first 10 min of the flood, and up to 6 times larger for the rest of the tidal cycle. The SSC peak was observed when the water depth was less than 0.30 m, whatever the hydrodynamic conditions. Suspended sediment transport during the rest of tidal cycle was also significant during stormy conditions but negligible during fair-weather conditions. 4.3. Sediment flux With the combination of the mean current velocity (Vmean) and the SSC, the near-bed sediment flux (Q) at 0.04 m (Micro-ADV and OBS) can be calculated. Its integration over the duration of the submergence period, gives an estimation of the sediment balance. During experiment 1 (moderate to stormy conditions), with a water depth of 0.51 m (during 24 min), a peak in sediment flux was observed (Qmax ¼ 1.1 kg/m2/s), followed by a noisy decrease from 0.4 to 0.1 kg/m2/s. This peak represented 56% of the total sediment flux during the experiment. At the end of the ebb, a weak peak attained 0.3 kg/m2/s while the water depth diminished from 0.16 to 0.04 m in 10 min (Fig. 4c); this peak represented 15% of the total sediment flux. Concerning this total sediment flux, two main sectors were observed: ESE (100e150
) (58% of the time) during the

SF EHWS MHW MHWN MLWN MLW MLWS

Fig. 2. Tidal range zonation at the experimental site (Point A). EHWS: exceptional high water springs, MHW: mean high water, MHWN: mean high water neaps, MLWN: mean low water neaps, MLW: mean low water, MLWS: mean low water springs; P: polder, SW: seawall, SM: Salt Marsh, SMF: Salt Marsh Fringe, SF: Sand flat.

a 13 11

9

7

5

3

347

water level from the Lowest Astronomical Tide (m)

R. Desguée et al. / Estuarine, Coastal and Shelf Science 94 (2011) 343e354

1

13

11

9

11/05/07

12/05/07

13/05/07

14/05/07

15/05/07

16/05/07

wind direction (°)

b

10/05/07

wind speed (m/s)

09/05/07

360 315 270 225

7

180 135

5

90 wind speed

3

45

wind direction 1 09/05/07

0 10/05/07

11/05/07

12/05/07

13/05/07

14/05/07

15/05/07

16/05/07

Fig. 3. Wind climate and water levels before and during experiment 1. Tidal curves on Fig. 3a are theoretical ones computed for Cancale (see location on Fig. 1), wind speed and direction were measured in Dinard (see location on Fig. 1) by Météo-France. The dotted box shows the period when experiment 1 took place.

flood period and WSW (240e260
) at the end of the submergence period (8%). Integration of the sediment flux over the total length of the tidal cycle yielded a sediment balance of about þ1790 kg/m2 towards the ESE (121.7
). During experiment 2 (fair-weather conditions), a peak in sediment flux (during 12 min) was observed with water depths of 0.04e0.33 m (Fig. 6c). Although slightly less important than for the moderate to stormy conditions of experiment 1, Qmax was, however, still significant (Qmax ¼ 0.87 kg/m2/s). Weak values close to zero were observed during the rest of the tidal cycle. The peak represented 83% of the total sediment flux. The two main sediment transport sectors observed were ESE (120e140
) for 25% of the tidal cycle, and NW (310e330
) for 39%. No peak at the end of the tidal cycle was observed. The sediment balance was about þ255 kg/m2 towards the South (173.7
). It is seven times less than that calculated during the turbulent conditions of experiment 1. 4.4. Bed shear stress (BSS) Bottom stress is a relevant parameter for analysing the deposition or erosion of sediment. However, the computation of shear

stresses is considered as complicated under extremely shallow water depths when classical laws are used (Le Hir et al., 2000). The results obtained in this paper concerning shear stress must therefore be considered as qualitative, but they do give a first approximation of shear stress conditions. Under moderate to stormy conditions (experiment 1), the waveinduced shear stress varied between 1 and 4 N/m2 (Fig. 7a). Most of the time, it was larger than the critical threshold for sediment movement defined by Migniot (1997) and Le Hir et al. (2008). Current-induced shear stress was always below the critical value obtained by Migniot (1997), which can be considered as a typical value for the sediments from the study area. Overall, the potential duration of local sand mobilization was about 101 min (82% of the submergence period), assured solely by waves. During fair-weather conditions (experiment 2), the wave shear stress was nearly constant (0.42e0.59 N/m2) and under the critical threshold for sediment movement (Fig. 7b). The current shear stress was at a maximum only at the beginning of the tidal cycle (1.57 N/m2) during a short period of about 6 min. Consequently, the potential duration of local sediment mobilization was very short, assured solely by tidal currents (6% of the submergence period).

R. Desguée et al. / Estuarine, Coastal and Shelf Science 94 (2011) 343e354

a

0.8 0.7 0.6 0.5 0.4 0.3 0.2

depth (m)/significant wave height (m)

348

water depth significant wave height HT

0.1 0.0

0.15

0.10

HT

0.05 current velocity suspended sediment concentration 0.00

suspended sediment concentration (kg/m3)

b 0.20

mean current velocity (m/s)

14/05/07 15:00 14/05/07 15:30 14/05/07 16:00 14/05/07 16:30 14/05/07 17:00 14/05/07 17:30

6 5 4 3 2 1 0

1.2 1.0 0.8 0.6

HT

sediment flux flux direction

0.4

flux direction (°)

c

sediment flux (kg/m²/s)

14/05/07 15:00 14/05/07 15:30 14/05/07 16:00 14/05/07 16:30 14/05/07 17:00 14/05/07 17:30

360 315 270 225 180 135 90

0.2 0.0

45 0

14/05/07 15:00 14/05/07 15:30 14/05/07 16:00 14/05/07 16:30 14/05/07 17:00 14/05/07 17:30 Fig. 4. Hydrodynamic conditions, suspended sediment concentrations and sediment flux during experiment 1. HT: high tide.

5. Discussion 5.1. Influence of wave activity on sediment transport on the upper tidal flat During the mild weather conditions of experiment 2, a typical “U”-shaped profile of mean current velocities was observed, with two asymmetric peaks, the stronger one at the beginning of the flood over a period of only 10 min, and the weaker one at the end of

the flood, with values between these peaks being near zero. This pattern is similar to that observed by Yang et al. (2007). This shape was also observed during other experiments with no waves present. Concerning the SSC, a typical “L”-shaped curve was observed during these experiments. A peak (2.5 kg/m3) occurred at the beginning of the flood, before a strong decrease up to the end of the tidal cycle. The weak concentration observed at the end of the ebb is the result of a sediment-deposition period around high tide when current speeds were close to zero. The curves obtained

a 13 11

9

7

5

3

349

water level from the Lowest Astronomical Tide (m)

R. Desguée et al. / Estuarine, Coastal and Shelf Science 94 (2011) 343e354

1

10

8

19/11/07

20/11/07

21/11/07

22/11/07

23/11/07

24/11/07

wind direction (°)

b 12

18/11/07

wind speed (m/s)

17/11/07

wind speed wind direction

360 315 270 225

6

180 135

4

90 2

45

0 17/11/07

0 18/11/07

19/11/07

20/11/07

21/11/07

22/11/07

23/11/07

24/11/07

Fig. 5. Wind climate and water levels before and during experiment 2. Tidal curves on Fig. 5a are theoretical ones computed for Cancale (see location on Fig. 1), wind speed and direction were measured in Dinard (see location on Fig. 1) by Météo-France. The dotted box shows the period when experiment 2 took place.

during experiment 1 are different. No peak in the mean flow velocity was really observed, but rather a noisy signal that was constant throughout the submergence period. Even though an SSC peak appeared at the beginning of the flood (6 kg/m3), values during the rest of the tidal cycle were not as negligible as during the calm conditions of experiment 2, especially at high tide. These field measurements show that local winds and windinduced waves are responsible for much of the sediment transport in the study area, as suggested by various authors working in other field sites (Ridderinkhof et al., 2000; Lee et al., 2004; Yang et al., 2007). A very high SSC of 6 kg/m3 was observed in the presence of wave activity. SSC values were also relatively high under negligible wind conditions, attaining up to 2.5 kg/m3 at the beginning of the local flood. The net total flux shows contrasts over the two meteorological periods presented in this study. The residual sediment transport during rough weather conditions was greater than that during fair-weather conditions by a factor 7. These high values suggest a large availability of sediments coming from the lower part of the tidal flat. Wave activity in bays or sheltered coastal environments often enhances sediment erosion both by increasing the bottom shear

stress induced by oscillatory flow and sometimes by fluidizing or liquefying the sediment through the generation of excess pore pressure on the bed, especially for fine sediments (Maa and Mehta, 1987; Wit and Kranenburg, 1997; Mehta and Parchure, 2000; Lambrechts et al., 2010). In shallow water, wave-induced liquefaction has been considered as the main process causing bed erosion under small waves (Lambrechts et al., 2010). Fluidized sediment put into suspension may then be easily transported by a tidal current. Flume experiments (Tzang and Ou, 2006) show also that fluidization is often observed with fine sand under monochromatic non-breaking waves or in the surf zone (Maeno et al., 1999). Bed fluidization of the tidal flat around the Mont-Saint-Michel Monastery is a common feature and such dangerous “quick sands”, warnings of which are sign-posted by the local authorities, pose a serious hazard to tourists and shellfish collectors on foot at low tide. While fluidization plays a significant role in increasing the sediment flux on the upper part of the intertidal zone, it is also important to underline the erosional role of waves on the concave bluffs of meandering tidal channels. Bluff retreat of meandering tidal channels under the impact of small waves is a locally very active process, enhanced by transport of the dislodged and

R. Desguée et al. / Estuarine, Coastal and Shelf Science 94 (2011) 343e354

a

0.7 0.6 0.5 0.4 0.3 0.2 0.1

depth (m)/siginficant wave height (m)

350

water depth significant wave height

HT

0.0

0.4

0.3

0.2

3.0 HT

current velocity suspended sediment concentration

0.1

suspended sediment concentration (kg/m3)

b 0.5

mean current velocity (m/s)

22/11/07 15:00 22/11/07 15:30 22/11/07 16:00 22/11/07 16:30 22/11/07 17:00 22/11/07 17:30

2.5 2.0 1.5 1.0 0.5

1.2 1.0 0.8 0.6

HT

0.4

sediment flux

flux direction (°)

c

sediment flux (kg/m²/s)

0.0 0.0 22/11/07 15:00 22/11/07 15:30 22/11/07 16:00 22/11/07 16:30 22/11/07 17:00 22/11/07 17:30

360 315 270 225

flux direction 180 135 90

0.2 0.0

45 0

22/11/07 15:00 22/11/07 15:30 22/11/07 16:00 22/11/07 16:30 22/11/07 17:00 22/11/07 17:30 Fig. 6. Hydrodynamic conditions, suspended sediment concentrations and sediment flux during experiment 2. HT: high tide.

suspended sediments by strong currents during the flood. Monthly retreat values of up to 35 m for the Couesnon channel close to Mont-Saint-Michel Monastery have been measured (Desguée, 2008). Sediment released by these processes is actively contributing to accretion of the neighbouring salt marsh area. The high suspended sediment concentrations also suggest potentially significant sediment mobility due to the specific characteristics of the local “tangue” (Larsonneur, 1994; Bonnot-Courtois et al., 2002). Although this mixed sediment may be classified as

non-cohesive sand-dominated (Van Ledden et al., 2004), the presence of small amounts of clay (2e5%) can affect the physical properties of the bed and its stability against erosion in particular. Clay particles may result in lubrication effect that can strongly reduce the critical shear stress (Barry et al., 2006), a conclusion also reached by earlier studies based on flume experiments (Torfs, 1995; Torfs et al., 2001). However, it has to be mentioned that experiments by Barry et al. (2006) were done for sand grain sizes (D50 between 0.41 mm and 1.20 mm) greater than those of the study

R. Desguée et al. / Estuarine, Coastal and Shelf Science 94 (2011) 343e354

4.5 4.0 3.5 3.0 2.5

current-induced

bed shear stress (N/m²)

a

351

wave-induced critical (Migniot, 1997)

HT

2.0 1.5 1.0 0.5 0.0

b 2.0

1.5

14/05/07 15:30

14/05/07 16:00

14/05/07 16:30

14/05/07 17:00

14/05/07 17:30

current-induced

bed shear stress (N/m²)

14/05/07 15:00

wave-induced HT

critical (Migniot, 1997)

1.0

0.5

0.0 22/11/07 15:00

22/11/07 15:30

22/11/07 16:00

22/11/07 16:30

22/11/07 17:00

22/11/07 17:30

Fig. 7. Shear stresses during experiment 1 (Fig. 7a) and experiment 2 (Fig. 7b). HT: high tide.

area (D50: 0.110 mm). More experimental studies are needed to confirm this effect over a large set of sand diameters and sand distributions and to characterise the clay content and its consequences on sediment transport in the area of interest. A final unknown quantity concerns the role of tidal flat macrofaunal species, which have been shown to alter bed stability, especially through the production of an easily erodible “biogenic fluff layer” from the bed matrix (e.g., Andersen et al., 2005; Lumborg et al., 2006). This layer comprises low-density, mucusenriched aggregates formed of faecal or pseudo-faecal pellets that are eroded before general bed failure (Orvain et al., 2007). During fair-weather, bed shear stress values may explain both the high SSC values at the beginning of the local flood (SSC close to 2 kg/m3) and the smaller values during the rest of the submergence period. This good relationship (Fig. 8) shows that even a low BSS can induce sediment transport, thus suggesting that the erosion threshold of natural sand/mud mixtures is lower than 1 N/m2 and that a value between 0.25 N/m2 and 0.50 N/m2, as obtained by Le Hir et al. (2008), may be more realistic. During experiment 1 (rough conditions), the current-induced shear stress was very low, always below the critical value, as was the wave-induced shear stress but only during a few minutes just at the beginning of the local flood. At this time of the tidal cycle, waves cannot induce the observed high turbidity values. Erosion is

commonly observed on the lower part of the flat and deposition on the upper flat. As indicated previously, under wave action, erosion observed on the lower parts of the flat is induced by oscillatory flow, and also probably by excess pore pressure. During the 5 days preceding the experiment, wave activity was noticeable with southwest winds of up to 10e12 m/s, thus probably inducing sediment resuspension. A wind rotation towards the northwest was also observed just at the beginning of the experiment. These wind and wave conditions seem to be responsible for transporting sediments towards the upper flat, even though tidal- and waveinduced currents were weak just at the beginning of the tidal flooding. Wave activity on the upper part of the flat also contributes directly to erosion as indicated by the wave-induced shear stress during this experiment. However, these periods of wave “efficiency” do not represent more than about 10% of the time on an annual basis. Assessing the contribution of local turbidity versus more general turbidity is difficult without measurements across the whole sand flat. Both can also explain the very high sediment concentrations at the beginning of the local flood especially under wave actions. Flats are also cut by tidal channels in which, as indicated above, strong steady currents can induce erosion, thus releasing sediments that may be transported towards the neighbouring flats during the flood tide.

352

R. Desguée et al. / Estuarine, Coastal and Shelf Science 94 (2011) 343e354

3

suspended sediment concentration (kg/m )

3.0

2.5

2.0

1.5 1.0

0.5

2nd order polynomial 2

regression (R =0.89)

2

current-induced bed shear stress (N/m ) 0.0 0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

Fig. 8. Suspended sediment concentration as a function of current-induced bed shear stress during experiment 2.

These experiments bring out the fact that antecedent meteorological conditions are determinant factors in suspended sediment concentrations and sediment transport on the tidal flat of MontSaint-Michel Bay. The stormy conditions observed during the days preceding the experiments were responsible for strong water mass turbidity, which moves with the tide towards the salt marshes during spring tides. Thus, by being responsible for local wave generation, wind intensity, direction and duration appear to be important factors controlling sediment fluxes in the western part of Mont-Saint-Michel Bay. 5.2. Wave-induced onshore sediment movement on the upper tidal flat Total sediment flux directions show a net transport towards the inner part of the bay that ultimately contributes to deposition over the adjacent salt marshes, even during rough weather conditions. The net sediment movement direction over flats has been commonly discussed in the literature. Generally, tidal processes generate onshore sediment movement and induce accretionary convex profiles, whereas wave activity results in erosion and concave profiles associated with offshore sediment transport (e.g., Kirby, 2000). In the present case, the onshore sediment movement observed during stormy conditions must be considered with regards to the experimental site. Le Hir et al. (2000) have suggested that waves effect is at its maximum at the beginning of the local flood, inducing sediment resuspension, the sediment being then transported onto the upper part of the flat. The asymmetrical orbital excursions near the bed on shallow tidal flats also reinforce onshore sediment transport (Lee et al., 2004). The tidal currents during the flood period, oriented towards the salt marshes, also contribute to the advection of suspended sediment. Permanent accretion of the tidal flat of Mont-Saint-Michel Bay is confirmed by results from sedimentation plates reported by Desguée (2008). Such accretion during rough weather conditions seems contrary to the erosion generally described in the literature (Bassoulet et al., 2000; Janssen-Stelder, 2000). This particularity suggests that sediment transport on very large tidal flats may show marked spatial contrasts. The net onshore sediment transport on the front part of the salt marsh contributes actively to the accretion of these marshes where tidal channel meandering does not occur, especially

during stormy conditions. This is corroborated by the permanent extension of salt marshes over the last decades (Desguée, 2008; Puissant et al., 2009). During calm conditions (experiment 2), the sediment flux is only significant at the beginning of the flood. It subsequently drops sharply, down to values near zero. As far as the sediment balance towards the salt marshes is concerned, its value is about þ307.5 kg/m2 during the first 9 min, and about 54.1 kg/m2 during the rest of the tidal cycle. During experiment 1, the sediment flux exhibits a peak at the beginning of the tidal cycle, but remains significant almost throughout the duration of the experiment, except at the end of the tidal cycle, in spite of a small peak in the tidal current as in the case of the fair-weather measurements. Clearly, during these rough weather conditions, the salt marshes and the upper part of the sand flat trap all the sediments brought in by the flood tide. Sediment transport was always oriented onshore attaining 365.9 kg/m2 during the first 9 min, a value close to that calculated for the same period during fair conditions, and þ1432.7 kg/m2 for the rest of the tidal cycle. No offshore movement (negative value) was observed for these conditions. These sediment transport data show that the beginning of the flood can thus be considered as the main period of prevalence of sedimentation processes on this higher part of the tidal flat. The results show that maximum sediment transport takes place during the first minutes of submergence at the beginning of the flood. During fair conditions, 83% of the total flux was observed during only 11% of the local tidal cycle duration (12 min) and with water depths between 0.04 and 0.33 m. During moderately stormy conditions, 56% of the total flux was measured just at the beginning of the flood with a water depth varying between 0.04 and 0.5 m, and 15% during 10 min at the end of the submergence period when the water depth fell from 0.16 to 0.04 m. In all, 71% of the sediment transport took place during about 28% of the local tidal cycle. This emphasizes the prevailing influence of shallow waters (