Variability of residual flow drivers in estuaries: numerical investigations

Residual exchange flows in estuaries can be either laterally sheared, vertically sheared, or a mix of both, depending on the competition between baroclinic and barotropic forcings. Observations of the residual flow structure at subtropical, semi-arid and temperate estuaries indicate that the dominant forcing (horizontal density gradient or tidal stress) may vary at a fortnightly to seasonal timescale. This work aims at characterizing the variability of residual flow drivers in estuaries based on process-oriented numerical simulations with simplified geometry. The hydrodynamic model (Delft3D) is constituted by a straight estuarine channel of 80 km in length and a Gaussian-shape cross-section of 1 km in width, forced by a semi-diurnal (M2) tide. The computational grid includes 240 and 15 nodes along and across the channel, respectively (about 333 m x 67 m resolution) and 20 uniform vertical layers. A total of 20 runs of 6 months were performed to explore the effects of various channel depths, river discharges and tidal amplitudes on the spatial structure of residual flows. The results show distinct cross-channel structures for each experiment and suggest a switch of the dominant driver (between tidally-driven and density-driven) along the channel in some cases. The spatial variability of residual flows is first examined against both the horizontal density gradient and the tidal velocity amplitude to characterize the competition between barotropic and baroclinic forcings. The dynamics of residual flow is then approximated by the tidally averaged momentum equation, applied at every grid point of the model domain, to derive dimensionless parameters (such as the densimetric tidal Froude number) for prediction of the dominant residual flow drivers both temporally and spatially (i.e., along the channel).