Due to climate change and human interventions, saltwater intrusion is becoming a topic of increasing concern worldwide. Salt water intrudes into the Rotterdam Waterway (RWW) by an exchange flow, where the denser sea water propagates landwards at the bottom. The main competing mechanism for this stratified exchange flow is vertical mixing, which can be realised by internal wave induced shear instabilities or wave breaking. The goal of this study is to investigate whether internal waves generated over undular bottom topography in the RWW can generate additional vertical mixing. The underlying assumption is that a decrease in stratification decreases salt intrusion.   The approach to answer the main research question is a combination of an analytical and a numerical analysis. The analytical study is based on frictionless linear theory. Internal wave behaviour is further analysed with FinLab, a finite element model which includes the non-hydrostatic processes and effects of density differences. FinLab is evaluated for the application of this study by means of a validation case.  In the analytical study, linear theory is applied to obtain a relation between the bed wave parameters and average internal wave energy density E for internal waves generated over sinusoidal bottom topography in a linearly stratified fluid. The derived expression describes that the bottom topography amplitude h0 and bed wave number kT both have a positive quadratic relation with the energy. Additionally, kTkinfluences the resonance conditions.  To validate FinLab for internal wave breaking and mixing an experiment in a wave tank, according to an example from literature, is simulated. The validation case reveals a shortcoming in the turbulent mixing parameterization. However, on scales relevant for the RWW the effect of this will not have the same significance. The validation case offers a suggestion for a subgrid closure of diffusion, where density effects are taken into account.  Numerical simulations of a 2D channel stretch with sinusoidal bottom topography, a linearly stratified fluid and a linearly varying background velocity, show generation of resonant trapped internal waves for the first two resonant modes. These occurrences correspond to the highest values of kinetic energy as function of vertical velocity averaged over the bed wave domain. The vertical buoyancy flux b is downward directed during occurrences of internal waves and becomes upward directed for increasing background flow. Vertical mixing is associated with an increase in average potential energy Ep, which is 17% higher for the base case (containing bed waves) than for a similar case without bed waves. This increase is larger when bottom shear stress increases. Richardson numbers below 0.25, associated with shear instabilities and mixing, are only observed near the bed, mainly when internal waves are present. The effect of variations in bottom topography wavelength LT and amplitude h0 on internal wave energy can be explained by the analytical formulation. The effect of bed wave parameter changes on b and relative increase in Ep can be related to the effect of the changed amount of bed friction rather than the difference in wave energy.   The first resonant mode is the most energetic, however, the average energy density found for these waves is only 0.4% to 6.7% of the potential energy anomaly (PEA); the energy required to fully mix a stratified water column. In the simulations the only mechanism that could transfer internal wave energy to turbulent kinetic energy are shear instabilities near the bed. Over the full simulation, the net vertical buoyancy transport is of negligible magnitude, where Ep shows significant increases between 6% and 99% compared to similar cases without bed waves and is enhanced during the presence of internal waves.  The main discussion point is that the quantification of vertical mixing requires improvement, particularly to determine the importance of mixing by internal wave-induced shear instabilities and by bed shear. Mixing by local shear instabilities (of which the relevant scales cannot be resolved with the current grid resolution) does not have an adequate parameterization, because density effects are not included in the turbulence closure. The bed friction parameter, which greatly influences the behaviour of the system, has to be validated. Furthermore, cases where internal waves might break in practice (e.g. at banks) were not considered. Finally, the observed internal wave energy is of small magnitude, however field measurements by Pietrzak(1991) shows that turbulence production by internal waves was significant.   

This study aims to get an insight into the particle trajectories that microplastics follow, after having been released in the North Sea.

For the computations daily-mean values of the surface currents are used, retrieved from the Mercator global ocean model. 2D particles trajectories are simulated for a year, with a 3rd party Python toolbox for Lagrangian simulation of particles: OceanParcels. Particles released from any location in the North Sea eventually get trapped in the Norwegian Coastal Current (NCC). From here they are being further advected to the North, at different moments in time for the particles released at different locations. The coastal processes in the NCC are mainly linked to wind and stratification, hence variations in ow patterns near the coast are linked to the seasons. When these ow pattern include large scale eddies, the particles follow a meandering and erratic path. Floating plastic particles released in the North Sea will flow northwards along the coast of Norway. Eventually those particles will end up in the Arctic region or get
trapped in the Norwegian fjords, independently of the location of release. However, the time scale of the northward advection depends both on where the particle has been released and the environmental conditions.

High Temperature Aquifer Thermal Energy Storage (HT-ATES) systems can be implemented to store geothermal energy, residual industrial heat and solar heat. Due to the large temperature differences between the injection fluid and ambient groundwater, buoyancy flow occurs, which causes heat losses which in turn results in poor efficiency of such systems. The heat losses can be reduced by applying density difference compensation, this entails injection water with a higher salinity, to compensate for the low density caused by the high temperature of the injection fluid.

The goal of this study is to get insight into the behaviour of a possible HT-ATES system on the campus of the TU Delft with the injection of saline groundwater from deeper layers for storage in both less deep and less saline aquifers. To achieve this different cases will be considered; a reference case, a theoretical optimum, density difference compensation of the TU Delft case and the optimum density difference compensation. For these cases, different scenarios are modelled in SEAWAT to identify the influence of injection temperature, injection volume, aquifer thickness, and hydraulic conductivity on the behaviour of the system.

Using the density difference compensation from the aquifer at TU Delft, with a salinity of 16.4 kg/m$^3$, a maximum improvement of 2.0\% can be achieved. Also, differences between the different scenarios for the TU Delft case were maximum 2.5\%, with exception for a thicker aquifer or smaller injection volume. Due to the large injection volume in thin aquifers, the hot water has a large volume, causing conduction losses to dominate in the system. For thick aquifers a lot of density driven flow occurs in the reference case. With optimum density difference compensation, having a salinity of 39.1 kg/m$^3$, the density driven flow can be significantly decreased and the efficiency increased with 21.5\%. Both aquifer thickness and salinity of the injection fluid have a negative correlation with salt recovery.

Thin aquifers in relation to injection volume are not sensitive to density driven flow, so density difference compensation is not useful for such HT-ATES systems. Density difference compensation in the case of the TU does give significant improvement in efficiency. The most influential parameter on the efficiency is aquifer thickness, for both the cold and warm well, in relation to injection volume. For thick aquifers with optimum density difference compensation the highest efficiency of 75.5\% is achieved after 10 cycles.