SEAWAY 



81 



factors. The reasonably indicated etfect of air-water 

 temperature difference should be included in the basic 

 data, rather than left to be a\eraged out. The effect of 

 limited water depth on long wave components may have 

 to be considered. Instrumental corrections may have 

 to be given more attention. 



The author doubts, howe\-er, whether a simple empiri- 

 cal data-fitting techniciue can be successful even with the 

 increased care suggested in the foregoing paragraph. A 

 semi-empirical approach based on the rational concept of 

 energy transfer from wind and dissipation of it in waves 

 may well give more nearly universally valid relationships. 

 The prespectrum work of Hverdrup and Alunk, Section 



5.1, was of this type and was outstandingly successful. 

 So also was the prespectrum work of Neumann, Section 



5.2. Reasonable energy-l:)alance equations can be set 

 up, leaving the values of certain coefficients and param- 

 eters to be derived from the observed sea data. Often 

 the expedient of grouping certain parameters permits a 

 better defined arrangement of observed material. 



While spectral deiinitions are a great advance in sea- 

 way description, they tend to concentrate attention on a 

 static concept of energy per unit area. The dynamic 

 (energy-flow) concept of the prespectrum work of Sver- 

 drup and Munk, and of Neumann should be blended 

 with the contemporary spectral-seaway concept. 



The semi-empirical approach suggested in the forego- 

 ing requires data gathering in a suitable form. Wave 

 spectra determined simultaneously for several positions 

 along the fetch will l)e needed. It will be difficult to 

 arrange such observations in the open ocean, and the 

 development of suitable spectrum formulations may de- 

 pend on small-scale experiments in limited water areas. 

 A method of scaling to larger ocean waves may be formu- 

 lated on a .semi-rational, semi-empirical basis. The 

 necessarily more limited obser\-ations in the open sea 

 may then be used either for verification or for evaluation 

 of isolated crucial parameters. 



9.2 Rational Approach. The word "rational" is rel- 

 ative. Esscnitially all research has to start from a cer- 

 tain empirical basis. There is, however, a rather defi- 

 nite distinction between semi-empirical and rational 

 approaches in their objectives. The semi-empirical ap- 

 proach can be thought of as a stop-gap; the objective is 

 to arrive at a workable solution as fiuickly as possible, 

 aiming at the best a\'ailable approximation to the na- 

 tural e\'ents but not necessarily seeking a profound 

 understanding of these events.'" 



In the rational methods, on the other hand, a sub- 

 stantial understanding of wave generation and wa\'e be- 

 havior is sought. There is no desire to arrive at a solu- 

 tion quickly at any cost, rather, emphasis is put on a 

 thorough investigation of the phenomena iinoh'ed and 

 on steadfast adherence to the principles of hydromechan- 

 ics. The results of such investigations are freer of un- 

 certainties of empirical observations. 



*'> The very u,soful pragmatic activity in .spectral representation 

 of the sea surface has been a regression in this respect from the pre- 

 sj)ectrum work of Sverdrup and Minik and of Neimiann. 



The phenomenon of wave growth under wind action is 

 a complicated one and research must l)e directed to un- 

 covering its many facets before the whole solution can be 

 formulated. Bowden's work cited in Section ;j.l and the 

 work outlined in .several sub.sections of Section 4 are on 

 isolated facets. The evident weakness of the current ac- 

 tivity is the lack of a comjirehensive guide which would 

 indicate the proper place ff)r all facets. While the valid- 

 ity of the results would still depend on improved knowl- 

 edge of individual facets, such a guide would be of great 

 value in organizing and channeling miscellaneous ac- 

 tivities. Apparently the only attempt in this direction 

 was made by Neumann in his prespectrum wcjrk. Section 

 5.2 and Appendix B. De.spite the fact that his study has 

 for its prime objecti\'e arrival at a directly usable semi- 

 empirical solution, it could serve well as a guide for the 

 new rational exposition. Its drawbacks are that it has 

 not been translated into English and that it is over- 

 burdened with detail considerations. 



One difficulty in developing a rational wave-growth 

 theory is described by the following free translation from 

 Neumann (195-4) attributed to A. Defant: "the unsatis- 

 factory condition appears here in that theoreticians apply 

 themselves to the advancement of theories, and with 

 completion of these their interest fades; investigators 

 working on the observational material, on the other 

 hand, have mostly too little theoretical training to apply 

 achievements of theory to particular cases. Close co- 

 operative developments of ob.servations and theory would 

 here, as everywhere, be very beneficial." Since it is ex- 

 tremely difficult to find an indi\'idual posse.ssing all re- 

 quired qualities, it appears that best progress can be 

 achieved by well-matched pairs or groups of investigators. 



It should not be assumed that "rational" research is 

 wholly theoretical, nor that it is limited to the use of 

 high-le\-el mathematics. The multiplicity of the facets 

 of the problem permits utilizing theoretical work of all 

 degrees of complexity, and calls for a wide \-ariety of ex- 

 periments and obser\'ations. 



Only one thing is imiversally re(|uired; namely, to be 

 at all times conscious of the ultimate use of the results of 

 a particular research, and to visualize how a particular 

 facet fits into the whole. This appears to have been 

 disregarded often in wind-flmiie experiments. The pri- 

 mary ol:)jecti\'e there has usuall.v been to e\'aluate the tan- 

 gential drag force of a wa\'y water surface, but this sur- 

 face has been only crudely described by an average wave 

 height, ignoring other wave-surface characteristics of the 

 wide variety which exists. In particular, in these experi- 

 ments the wind speed-to-wave celerity I'atio, U/c, is 

 usually too high and the experimental wa\'e steepness 

 H/X is also too high in comparison with the natural large- 

 scale phenomena. 



The empirically obtained spectra formulations, the 

 theoretical work on spectrum properties by Pierson, 

 Longuet-Higgins and Cartwright, and the theoretical 

 work on wave generation by Munk, Miles, and Phillips 

 indicate that the results of past work in wind flumes and 

 in limited water areas are of little u,sefulness. E.s.sentially 



