SECT. 5] WIND WAVES 675 



Work on similar lines is in progress in other countries, but accounts are not 

 yet published. 



There is some doubt as to whether the loss of energy from wave trains by 

 viscosity has to be taken into account. The theoretical difficulties have been 

 referred to in section 3, page 671, and practical measurements are inconclusive. 

 Darbyshire (1959), and Pierson, Neumann and James (1955) consider only 

 the spreading of energy, whereas Gelci takes into account energy loss due to 

 turbulent viscosity. 



There is some evidence that the temperature difference between the water 

 and the air has an appreciable effect on wave generation. If the air is colder 

 than the sea, it is unstable and the waves might be expected to be larger for a 

 given wind speed. Fleagle (1956) finds that in these circumstances the wave 

 height increases 10% per °C temperature difference. M. Darbyshire (1958) also 

 finds an appreciable effect in some circumstances but J. Darbyshire (1959) 

 finds it to be undetectable. The effect is not sufficiently well-established for it 

 to be used normally in wave prediction. 



A. Accuracy of Wave Predictions 



Though some practical methods of wave prediction are based on formulae 

 derived partly theoretically, the theory is in all cases unsatisfactory, and the 

 formulae can only be relied upon to the extent to which experience has shown 

 them to be reliable. Perhaps the most graphic way of demonstrating the sort 

 of accuracy obtainable is to compare some predictions for simple conditions. 



Table II (figures from Francis, 1959) gives forecasted wave heights for a fetch 

 of 50 nautical miles over which the wind has been blowing for a long time. 



Table II 



This comparison is not, perhaps, entirely fair, since the methods are usually 

 used for longer fetches. They are supposed to apply to fetches of 50 miles, 

 however, and the agreement can hardly be considered to be satisfactory. 



Pierson (1959) has compared predicted with measured spectra. He finds that 

 the spectrum predicted by the Neumann formula does not agree well with that 

 observed. 



