796 



[chap. 23 



These results, obtained by difference methods, in one dimensional (tidal 

 estuaries) as well as in two dimensional cases (North Sea) give an idea of the 

 precision obtained by these numerical methods. Now that this preliminary 

 work has been done, the problem of oceanic tides, especially in the Atlantic 

 Ocean, ought to be tackled by means of these difference methods. Computations 

 with relatively large mesh-size have shown that in suitable coastal areas with 

 small width of continental shelf there is, to some extent, agreement with 

 observations. In areas where the shelf is wide these differences apparently 

 increase ; this suggests that the depth variation at the edge of a continental 



HI q[ ^=^=:'. 



Fig. 27. Tides and storm-surge in the North Sea, January, 1954. Full lines and open circles, 

 observations at gauges in Lowestoft L, Harwich H, Ijmuiden I, Terschelling T, 

 Cuxhaven C, Biisum B, Esbjerg E, Hirtshal« Hi ; broken lines and dots, computation. 



shelf is important. As previously mentioned, the difference method is not only 

 applicable to tidal problems but also to more general problems of dynamical 

 oceanography. External forces and boundary values are allowed to be arbitrary 

 functions of time and space. In this way it becomes possible to investigate the 

 interaction of tides and storm-surges. Fig. 27 contains the records of sea-level 

 at the gauges in Lowestoft, Harwich, Ijmuiden, Terschelling, Cuxhaven, Biisum, 

 Esbjerg and Hirtshals during the storm of January, 1954 (full lines) and the 

 computed values (dashed lines). The boundaries of the area as well as the wind 

 stress acting at that time are also shown. 



