674 



BARBER AND TUCKER 



[chap. 19 



It should be noted here that the "surface" wind speed V is defined as the 

 wind speed at a height of 10 m above the mean sea surface, but that there are 

 severe practical difficulties in its measurement at sea owing to the interference 

 due to the presence of the ship. 



By including other observations at short fetches, he finds that he can extend 

 this formula to cover all cases in which the water is deep compared with the 

 wavelength of the waves. 



It is more difficult to predict waves near the coast, since the problem is 

 complicated by refraction due to the changing depth of water and due to 

 tidal streams : absorption of wave energy may also be important. It has been 

 found possible to develop formulae for locations where many wave records 

 are available, but these are strictly local in application, and prediction of waves 

 for a location where no previous records have been taken is still far from 

 satisfactory. 



-0.04 



0.04 



0.08 



-0.04 



0.04 0.08 



f - {q, sec"' 



Fig. 4. Observed wave spectra for different wind speeds. Spectral density Hf^ divided by- 

 total energy H^ in spectrum is plotted against frequency deviation from /q, the 

 frequency of maximum Hf^. (After Darbyshire, Deut. Hydrog. Z., 1959. Figs. 10, 11 

 and 12.) Left-hand diagram: winds of force 3, 4 and 5. Centre diagram: winds of force 

 6 and 7. Right-hand diagram: winds of force 8, 9 and 10. 



So far, waves generated only by local wind have been discussed. This is not 

 always adequate for prediction, since swell from distant storms may be im- 

 portant. Lebel and Gelci (1959) have developed a method of wave prediction 

 which considers all possible sources of waves, and which produces a chart of 

 wave height covering the whole of the relevant area of ocean ; in their case, they 

 have considered the North Atlantic. They divide the ocean by a grid containing 

 839 meshes, and for each mesh they calculate the influx and effiux of energy 

 from and to adjacent meshes and the energy fed to the waves by the wind. The 

 directional energy spectrum is split into finite intervals of period and direction, 

 and the computation is made for each of these components. The computation 

 is repeated at 6-hourly intervals, and requires the results of the previous 

 computation. An electronic calculating machine is used. 



