Morphodynamic processes in shallow estuaries: influence of tidal flats and channels on sand transport

Flood/ebb-dominance in net sediment transport of shallow sandy estuaries with tidal flats and channels (typical around the U.K.) is investigated using a 2D model (TELEMAC). The micro-tidal Dyfi Estuary is used as a case study. The along-estuary net sediment transport in the Dyfi has been predicted and characterised as ebbor flood-dominant. The simulations illustrated that shallow waters (inner-estuary) give rise to flood-dominance, while flats and deep channels (outer-estuary) causes ebb-dominance. For a more generic classification, idealised estuary shapes (with/without tidal flats) have been modelled. For medium sands and no flats, the net transport switched from ebbto flood-dominant when a/h (tidal amplitude/water depth) > 1.4. A similar pattern, with greater transport, was simulated with flats included and also with reduced grain size. The limiting values of a/h are greater than those (based on tidal-asymmetry, used as a proxy for bedevolution) proposed by Friedrichs and Aubrey (1988) for deeper estuaries. level, MSL). The advantage of this parameter is that it does not require a modelling approach. According to Friedrichs and Aubrey (1988), ebb-asymmetry occurs when a/h < 0.2 and flood-asymmetry when a/h > 0.3. Their research, however, was based on sea surface variations and not the velocities. They did not incorporate sediment transport into their model to verify the results. Also, their classification is based on averages over an entire estuary. In reality, some parts of an estuary may experience floodasymmetry whereas others ebb-asymmetry. The Dyfi Estuary has been investigated here in order to understand the sediment transport and morphology of shallow estuaries with extensive tidal flats and deep channels. The estuary is located in Cardigan Bay on the west coast of Wales, U.K. (Figure 1). The estuary is constrained to a width of 1 km by Ynyslas Spit at the mouth and extends 8 km to the east; bounded to the south by the Borth-Dyfi Junction rail line. Afon Dyfi and Afon Leri are the two primary rivers that flow into the estuary and have been incorporated into the model. Their mean annual flowrates are 25 m3 s -1 and 5 m3 s -1 , respectively (Robins, 2008a). The estuary consists of medium sands (2.5×10 -4 m), and has extensive tidal flats that are exposed at mid-tide, in contrast to the meandering, deeper channel (at high spring tide, water depths in the channel reach 7 m, reducing to 2 m on the flats). Local waters are generally well-mixed in terms of density stratification and cover an estuary surface area of approximately 17.3 km3 during extreme spring tides which have a 4.9 m range (Shi, 1993). The estuary is currently ‘Type II’ in Pethick’s (1994) temporal classification of estuary development. Type I estuaries represent rapid infill after the Holocene transgression. Type II estuaries involve a channel/flats system leading, for the reasons stated above, to ebb-dominance. However, this classification is rather broad as an estuary may experience both types of dominance at different locations. Brown (2008), for example, simulated ebbdominance in the lower Dyfi Estuary and flooddominance further up-estuary. During extreme tidal, surge and fluvial events, Cors Fochno (Borth Bog), a low-lying nature reserve and Site of Special Scientific Interest (SSSI) and the tourist village of Borth are at risk from flooding (Robins, 2008a,b). Therefore, the estuary may undergo alterations in bathymetry and intertidal storage. It is evidently important to quantify and understand the processes which contribute to sediment transport in the Dyfi Estuary leading to generic conclusions for estuaries of this type. The TELEMAC Modelling System, described in Section 2, has been used to simulate the sediment transport in the Dyfi Estuary (Section 3). Subsequent simulations using idealised estuary shapes have then been performed (Section 4) so that the transport patterns could be explained more generically. A primary aim of this paper is to critically asses Freidrichs and Aubrey’s (1988) estuary classification parameter a/h using a shallow water case study. Finally, the results are summarised in Section 5. Figure 1. (a) The Dyfi Estuary (boxed), Wales, U.K. (b) The Dyfi bathymetry (from a model simulation). Borth Bog is the low-lying area south of the rail embankment. The dotted line through the mouth shows the transect plotted in Figure 3(c). 2. TELEMAC MODELLING SYSTEM The TELEMAC Modelling System (Hervouet, 2007) is well suited to modelling coastal and fluvial dynamics due to the finite-element grid allowing graded mesh resolution. This means that the model grid can achieve high resolution of nearshore processes and coarser resolution in deeper water which optimises model accuracy and computation time. In contrast, traditional finite-difference models can experience abrupt changes in grid element size when applied to the nearshore zone or steep bathymetry. The model, used here in 2D, vertically averaged mode (TELEMAC-2D), is based on the shallow water Saint-Venant equations of momentum and continuity, derived from the Navier-Stokes equations (see Hervouet, 2007). One of the restrictive hypotheses of these equations is that the wavelength should be large in relation to the depth (e.g. nearshore environments). The model has been forced here with tidal elevations and river flowrates. SISYPHE is the sediment transport module coupled with TELEMAC-2D to produce simulations of bed evolution. The Soulsby-Van Rijn transport formula (Soulsby, 1997) was used here where total (bed load + suspended load) sediment transport rate per width of the flow in combined waves and currents on horizontal and sloping beds is calculated. The transport rate formula is expressed: