Search

The dynamics of cooling water spreading in a non-tidal embayment is subject of a modelling-based study of Greifswald Bay, a shallow embayment at the south-western coast of the Baltic Sea. Potential cooling water spreading due to a possible power plant at Greifswald Bay is evaluated as differences between a realistic hind-cast simulation and a similar simulation but including the cooling water pumping. The model results are confirmed with satellite imagery of the embayment during operation of a nuclear power plant in the 1980s. The effect of cooling water pumping on the residual circulation, additional stratification and the heating of near-bed waters in the herring spawning areas is evaluated from the simulation. The model results for an idealised embayment and the realistic scenario, as well as the satellite images, show a clear dependence of the plume spreading on the wind direction. Although the surface plume affects a large area of the embayment, the results show a localised impact on residual circulation, bulk stratification and heating of the waterbody.

Observational evidence is presented, which corroborates the hypothesis of the general presence of estuarine circulation in the Wadden Sea as put forward in a previous study ( Burchard et al., 2008 ). Current velocity data from moored ADCPs (in the Hornum Deep south of Sylt Island, 2002–2009) and ship cruises (in several locations in the German Wadden Sea, 2000–2008) were analysed. As a general result, the vertical current profiles above the benthic boundary layer are usually more homogeneous during flood than during ebb, with a pronounced dependence on the cross-shore horizontal density difference. This tidal asymmetry consequently must lead to a residual outflow of Wadden Sea waters in the upper part and a residual inflow of water in the lower part of the water column, thus giving a generic explanation for the obvious net import of suspended sediments from the German Bight into the Wadden Sea.

Investigations on stratifying and de-stratifying processes in the Limfjord in Denmark, stretching from the North Sea at the western end to the Kattegat at the eastern end, were carried out. The forcing of the Limfjord is dominated by mean sea surface elevation and salinity gradients, diffuse freshwater supply, as well as episodic saline inflow events from the North Sea boundary. The physical regime of the Limfjord switches periodically from stratified to mixed conditions. In order to investigate the complexity of the system, a realistic three-dimensional high-resolution numerical model study was conducted for the year 2003. The results are based on, and compared to, available field data. A central subject of this study is the anomaly of potential energy as a measure for stratification. The stratifying and de-stratifying processes are identified as contributing terms in a dynamic equation for the potential energy anomaly. The comparison of model results and observations shows that the model has the capability of reproducing the observed physical regimes. Changes in the simulated stratification turn out to be a complex interaction of differential advection, heating and turbulent mixing in the central Limfjord.

Only recently, medium intensity inflow events into the Baltic Sea have gained more awareness because of their potential to ventilate intermediate layers in the Southern Baltic Sea basins. With the present high-resolution model study of the Western Baltic Sea a first attempt is made to obtain model based realistic estimates of turbulent mixing in this area where dense bottom currents resulting from medium intensity inflow events are weakened by turbulent entrainment. The numerical model simulation which is carried out using the General Estuarine Transport Model (GETM) during nine months in 2003 and 2004 is first validated by means of three automatic stations at the Drogden and Darss Sills and in the Arkona Sea. In order to obtain good agreement between observations and model results, the 0.5 × 0.5 nautical mile bathymetry had to be adjusted in order to account for the fact that even at that scale many relevant topographic features are not resolved. Current velocity, salinity and turbulence observations during a medium intensity inflow event through the Oresund are then compared to the model results. Given the general problems of point to point comparisons between observations and model simulations, the agreement is fairly good with the characteristic features of the inflow event well represented by the model simulations. Two different bulk measures for mixing activity are then introduced, the vertically integrated decay of salinity variance, which is equal to the production of micro-scale salinity variance, and the vertically integrated turbulent salt flux, which is related to an increase of potential energy due to vertical mixing of stably stratified flow. Both measures give qualitatively similar results and identify the Drogden and Darss Sills as well as the Bornholm Channel as mixing hot spots. Further regions of strong mixing are the dense bottom current pathways from these sills into the Arkona Sea, areas around Kriegers Flak (a shoal in the western Arkona Sea) and north–west of the island of Rugen.

In January/February 2004 a ship cruise into the Arkona Sea (Western Baltic) happened to start just prior to a medium-size inflow event and to last until a few days after the inflow. The bottom layer of saline inflow water in the southern Sound has a current speed of up to 1 m/s. The current bifurcates at Kriegers Shoal, where only the smaller branch follows the contour lines of the bathymetry at the western rim of the basin, while the main flow proceeds on the steepest path north of Kriegers Shoal. The pycnocline tilt due to geostrophy matches with current observations in both branches. Old bottom water in the basin is lifted up by the denser inflow and pushed forward to Bornholm Gate. Inflow from the Darss Sill at first consists of low-salinity Baltic surface water and a small amount of higher salinity water at the bottom. After a few days the water over the sill is mixed and moves east, establishing a strong subsurface current at the southern rim of the Arkona Basin. Neither the inflow through the Sound nor the water masses from the Darss Sill reached the Bornholm Gate in the east of the Arkona Sea before the end of the survey. The saline outflow observed there consists of bottom water that had been present in the Arkona Sea prior to the new inflow. Overflow water spreading on the sea bottom is modified by entrainment from the low salinity water body above. Best guesses of the entrainment rate, somewhat tentative due to salinity variations at the entrance, are in line with formulae from the literature.

Recent contributions to second-order turbulence modelling are reviewed with an emphasis on models of the coastal ocean. Classical and recent turbulence models are reformulated using a unifying notation, making differences in model structure and model parameters more transparent. The essential characteristics of a number of models are compared. It is further demonstrated that the mathematical stability and realisability of the models is crucial for practical applications. The parameter constraints to insure stable computations are discussed. Model performance is evaluated for a number of idealised entrainment scenarios that are typical for shelf seas: entrainment in linearly stratified and two-layer fluids caused by (a) surface wind stress (b) bottom stress due to water motion driven by a barotropic pressure gradient (c) bottom stress due to water motion in a bottom jet. It is shown that the predicted entrainment velocities are in good agreement with available laboratory data, provided the models have been calibrated according to a theoretical constraint. In addition, model behaviour in free convection is re-analysed with the help of recent data from large eddy simulations (LES). It is shown that even the most simple models, which do not solve a differential equation for the turbulent kinetic energy (TKE), can approximately reproduce the entrainment depth, however, they fail completely in reproducing the correct budget of the TKE-balance. In contrast, second-order models solving transport equations for the TKE and a length scale determining variable are in reasonable agreement with the TKE-budgets from LES.

The aim of this study is to intervalidate observations and numerical simulation results for the turbulent dissipation rate under strong wind conditions in the Northern North Sea during one week in October 1998. The observations were obtained by spatially and temporally averaging measurements of small-scale shear with a free-falling shear probe. The 1D numerical model used for this study is based on a state-of-the-art two-equation k − e turbulence model with an algebraic second-moment closure scheme. It is discussed by means of annual and seasonal model simulations how the influence of heat and salt advection and internal waves can be accounted for. After these precautions, the agreement between observations and simulations of the turbulent dissipation rate are fairly good. Remaining differences cannot only be explained by problems such as undersampling and noise level, but also by idealising model assumptions.

The Liverpool Bay Region of Freshwater Influence in the Irish Sea exhibits strong horizontal gradients which interact with the dominant tidal flow. A 25 h series of measurements of the cycle of turbulent dissipation with the FLY dissipation profiler shows a strong asymmetry between ebb and flood which is associated with a cycle of increasing stratification on the ebb and progressive mixing on the flood which results in vertical homogeneity as high water is approached. At this time strong dissipation extends throughout the water column in contrast to the ebb when there is a near shutdown of dissipation in the upper half of the column. The cycle of stratification and dissipation is closely consistent for the two semi-diurnal tidal cycles observed. We have attempted to simulate this situation, which involves a complex suite of processes including tidal straining and mixing, using a version of the k–e closure scheme in a 1-d dynamical model which is forced by a combination of the observed tidal flow and horizontal temperature and salinity gradients. The latter were measured directly at the end of the observational series but, in order to focus on the cycle of dissipation, the correct reproduction of the temperature and salinity cycle can be assured by a nudging procedure which obliges the model temperature and salinity values to track the observations. With or without this procedure, the model gives a reasonable account of the dissipation and its asymmetric behaviour on ebb and flood although nudging improves the timing of peak dissipation in the upper part of the water column near highwater. The model has also been used to examine the ratio of shear production (P/e) and buoyancy inputs to dissipation (B/e). The variation of these quantities over the tidal cycle confirms the important role of convective motions forced by tidal straining near the end of the flood phase of the tide.