Baltic Sea with Numerical Models

A refined one-dimensional numerical ocean model of the southern Baltic Sea was used by Axell (2002) to investigate suitable parameterizations of unresolved turbulence and compared it with available observations. The turbulence model is a k- model that includes extra source terms of turbulent kinetic energy production by internal waves and Langmuir circulation due to unresolved, breaking internal waves and Langmuir... 

The second part of the chapter deals with tides in the Baltic Sea, a much neglected subject in the author s opinion. The topic is introduced by a brief historic summary of Baltic Sea tidal theory, followed by an outline of present tidal theory for the Baltic Sea. The conclusion at the end of this chapter is that inclusion of the tides in numerical models of the Baltic Sea will lead to more realistic sea level forecasts. 

The quality of numerical simulations is limited by the horizontal and vertical model resolution. An obvious aspect of enhanced model resolution is the better representation of the bathymetry. Especially the structured coastline of the Danish Straits and the sills between the basins of the Baltic Sea are represented more accurately. Another aspect is the representation of processes with small characteristic dynamic scales in the model. Although in many cases details on unresolved processes are not needed, parameterizations for such processes are sources of numerical uncertainty. Enhanced model resolution means less parameterization and the model dynamics is represented preferably by the hydrodynamic equations. 

As described in Section 19.2.3.8, regional models, such as a Baltic Sea model, can get the boundary values for the calculation of surface fluxes from simulations with atmosphere models, which have been carried out previously. This is possible, because the influence of the Baltic Sea on the Northern Hemisphere weather system is only important for local phenomena, and inaccuracy in the feedback from the Baltic Sea to the atmosphere is of minor importance, Schrumm and Backhaus (1999). Widely used datasets, such as the ERA-40 reanalysis data, are improved by assimilation of observations. If surface variables calculated by the ocean model tend to drift away, this is compensated to a large extent by the calculated surface fluxes. For this reason numerical simulations with standalone ocean-ice models can be successful. 

---