Finite element modeling of tidal flow and bed -load transport in the Great Bay Estuarine system, New Hampshire

Current, sea level and bed-load transport are investigated in the Great Bay Estuary, New Hampshire---a shallow, well-mixed system with low freshwater input. Main channel tidal currents are over 2 m/sec, and its dynamic balance is dominated by the pressure gradient and bottom friction. Current and sea level forced by the M2, M4, M6 tides at the estuary mouth are simulated by two numerical models---a three-dimensional, harmonic model that solves the linearized shallow water equations (FUNDY5); and a vertically averaged, time stepping, non-linear model (ADAM). ADAM makes use of a kinematic assumption (that is, the local and advective accelerations are neglected) and accounts for flooding and dewatering over tidal flats by employing a groundwater component. FUNDY5 is used as a preliminary diagnostic tool to identify the general properties of the estuarine flow and to identify grid problems. ADAM is used as a prognostic model to simulate the tidal hydrodynamics of the estuarine system. The accuracy of the hydrodynamic predictions is evaluated by comparison with ten tidal elevation and four cross-section averaged current measurements. The results show that the kinematic assumption holds for the lower and middle sections of the estuary but fails in the upper sections, due in part to the increased importance of accelerations in these sections. Residuals (time averages) of currents are also investigated and yielded flood directed pathways over the shallow flats and ebb directed pathways in the deep channels. Currents simulated by ADAM are then used to model bed-load transport in the vicinity of a rapidly growing shoal located in the main channel of the lower system. Consisting of coarse sediments, the shoal must be dredged every 5--9 years. Two approaches are taken---an Eulerian parametric method in which elemental bed-load flux vectors are calculated at each time step; and a Lagrangian particle tracking approach in which a finite number of sediment particles are released at different times in the M2 cycle and tracked. Both methods yield pathways and accumulations in agreement with the observed shoal formation and the long-term rate of sediment accumulation in the shoal area