34 
The BSH models use a finite difference method and nested grids with constant grid spacing 
in the individual grids. Fig. 6.1.3 shows the effect of a grid spacing reduced from about 10 km 
to 2 km in simulations with two different two-dimensional, barotropic model versions (see 
sections 7.5 and 7.4). As the two models cover slightly different areas and the simulations 
were not started at exactly the same time, the time axis in Figs. 6.1.4 and 6.1.5 was shifted in 
such a way that the signals arrived at Wick simultaneously. The input signal (three positive 
signals, period 30 minutes, wave height 5 m) in the coarse-resolution model (Fig. 6.1.4, 
green) hardly shows any modification at Wick and is higher in the first two signals than in the 
fine-resolution grid (Fig. 6.1.4, red). In both models, Cuxhaven is reached first by extensions 
of the input signal that has been attenuated on the shelf, followed by a higher signal caused 
by diffraction (cf. sections 7.4 and 7.5). Flowever, the simulations differ in two important 
aspects. In the coarse-resolution simulation (Fig. 6.1.5, green), the primary signal is higher, 
and the secondary signal lower, than in the simulation using the finer grid (Fig. 6.1.5, red). 
Besides, in the coarse-resolution model the primary signal reaches Cuxhaven about two 
hours later; the secondary maximum occurs about one hour later. Since not only a refined 
grid but also an improved bottom topography was used in the North Sea model with the 2-km 
resolution, any conclusions as to the dependence of propagation velocity on the grid 
resolution would be premature. 
Fig. 6.1.3: Water level distribution with equal input signals (3 positive single signals, period 
1800 s, wave height 5 m) and different models. Left: North Sea model 10 km 
resolution (after 8 hours), right: North Sea model 2 km resolution (after 9 hours).