|
[menu.htm]
|
A European Commission DGXII MAST 3 Project |
|
Coastal Wave Modelling
Once the offshore information is available, and after the user has indicated on the screen the area and the specific point P (demo of how this is done) where he requires the results, the transfer of the wave conditions to P may be performed. This involves the following steps, each without any intervention by the user:
- A fine local grid is established;
- The most suitable wave model is selected and run;
- Statistics for P are calculated (this last point will be discussed further in the next section. Here we deal with points 1 and 2).
Establishing the grid
In Eurowaves, it is planned to provide results at least as close as five kilometres from the coast and down to five metres depth. This is basically connected to the practical accuracy with which we know the bathymetry close to the whole European coastline. If, as mentioned in the previous section, more detailed local information are introduced, the user can move even closer to the coast.
However, a practical difficulty may be the computer time required for the calculations. This will be discussed later. The point is that, as we move closer to coast, the grid must have enough resolution to describe the important local details. On the other hand the grid must extend offshore at least to the points where the wave conditions are known (see the second section) in order to provide the necessary boundary conditions. A finer grid will lead to a rapid increase in the number of points, and hence of the time required for the calculations.
The time will depend in a critical way also on the type of model selected by the user. Eurowaves will offer several models to the user, ranging from the simple backtracing ray technique which considers only refraction, shoaling and, possibly, generation by wind, to the advanced third generation models that take full consideration of all the relevant processes that govern wave evolution in both deep and shallow water. As a rule, the more sophisticated models should produce better results, at least in complicated bathymetric areas, very close to the coast and for extreme wave conditions. Therefore, the choice will often be a trade off between the resolution at the coast, the accuracy of the results and the time required for the calculations. A possible optimum strategy for the user could be to attack the problem first with a simple (and fast) model, followed, possibly, by a more elaborate approach for the final calculations and/or for analysis of special events such as extreme storms.
Wave modelling
It is clear that the user must be given suitable information on the possible accuracy of the different models. For this purpose a lot of effort in Eurowaves is being put on validation and we are presently testing the various models against measured data. For this purpose, multiple measurements are required. Given a suitable location with certain characteristics, e.g. a bay with a sloping bottom, we need contemporary wave data taken offshore, say at A, and at a more inshore location B. Using the A data as input, we can model the transfer to B and verify the model results versus the data from B.
One of the most valuable of such data sets being used for this purpose has been collected at Holderness, on the East Coast of England. Prandle et al. (1996) give a full description of the measurement campaigns and the available data. The transfer has been modelled so far with the SWAN model (Booij et al., 1996), http://swan.ct.tudelft.nl/ a recently developed third generation model, specially designed for shallow water, and with VENICE (Cavaleri and Malanotte Rizzoli, 1981), a back-ray tracing model. A sample of the results from the SWAN simulations (the number of runs of the model amounts to several tens of thousands) is given in the figure. The A and B buoys mentioned above (N3 and N1 in the figure) were moored at about 12 and 1.6 kilometres off the coast, at respectively 30 and 12 metres depth. Among other possibilities, we explored 1) the effect of having as input a two-dimensional spectrum or only the summarising parameters (H, T, q ); 2) the influence of the local wind on the results. We note from panels a) and b) that the parametric approach leads to higher values at the inner buoys with respect to the spectral approach. The reason is that, considering only the cases with dominant waves moving towards the shore, and summarising the wave conditions at A (N3) by a single triplet, we include inherently into the model onshore propagating input also the wave components which were actually moving offshore.
A similar argument applies to the effect of the wind (panels c and d in the figure). A wind towards the shore leads generally to higher waves closer to the coast, hence at B (N1). However, this can also occur with the model even with an offshore blowing wind. For the same reason as that explained above, if the dominant wave system is moving onshore, the buoy will add up the two systems into the triplet (H, T, q ), providing an apparent input to shore larger than the actual.
Sclavo and Cavaleri (1999) give a fuller discussion on this. It is clear that these effects need careful consideration when deriving and interpreting coastal statistics from a wave model. From the WAM model, we have 6 parameters, 3 each for wind sea and swell. Using all 6 parameters will potentially solve this problem. Work is proceeding with similar model validation work against other data sets in Europe where buoy data are available in offshore and coastal waters simultaneously (e.g., Crete, Portugal, Spain, Italy and Norway).
Figure. Wave model simulations of the data collected at Holderness, UK. N3 offshore buoy, N1 inshore buoy. Top, left and right: comparison between the spectral and the parametric approach; bottom, left and right: comparison between the measured and modelled wave height, taking wind into account.
This site is best viewed with Internet Explorer 4+ or Netscape Navigator 4+. Javascript must be enabled for the navigation menu to operate. Please note there may be linked pages also under menu items having sub-menus.