SEAWAY 



51 



o 



to 



V-24 Kn 



« • I <| . * - A 



Neumann 



\ 



♦ l*»*C« r* * • 



_L 



15 20 



Windwirkungsdauer 



25 



30 



35 S+d. 



Fig. 56 Apparent wave periods plotted versus 24-knot wind duration (abscissa scale in hours). Dots indicate weather 

 ship reports made in intervals of 2 sec (from W'alden, 1953/54) 



methods, Braceliu's (prespectrum) one gives results in 

 best ugreenuMit with Waldens' anah'sis of weather-ship 

 observations for wind duration up to .'-iO lir, I'"or longer 

 duration it somewhat exaggerates the wa\'e heights. 

 Neumann's method gives a too slow wave rise diu'ing the 

 first 5 hr of wind action, but strongly exaggerates waves 

 after that. 



Fig. 56 compares the apparent wave periods computed 

 by Neumann's method with those reported liy weather 

 ships. The apparent periods are reported by ships in 

 2-sec intervals, and the resultant series of points is seen 

 to straddle Xeumann's period when wind duration ex- 

 ceeds 15 hr for a 24-knot wind. This finding is in agree- 

 ment with results demonstrated by Neumann in Fig. 

 41. It can be considered as partial confirmation of the 

 shape of the Neumann spectrum. Tlie prediction of the 

 wave height, indicated by this spectrum, can be im- 

 proved following suggestions by Walden . On the basis of 

 his data Walden (1953/54) recommended reduction of the 

 constant in Xeumann's spectrum expre.ssion. Later, 

 Walden (l()5fi) recommended a reduction of this constant 

 on the basis of a reinterpretation of Neumann's Fig. 36. 

 The apparently too slow initial rise of wa\'es shown by 

 Neumann's method remains uncorrected. 



6.53 Practical wave prediction. In the ca.ses cho.sen 

 by Walden in the preceding section, weather conditions 

 had to correspond to the assumptions made for the 

 spectrum formulations. In normal wave forecasting, 

 however, the effects of fetches with different wind 

 velocities, of moving fetches, ^^ of .several fetches together, 



» Reference.s, Walden (195G6), Wilson (1955) 



of swells from distant storms, and so on, must be con- 

 sidered. It almost appears that the initial spectrum 

 formulation becomes subordinate to the forecaster's 

 reading and interpretation of weather maps. Neverthe- 

 less, the particular formulation is important since a 

 foreca.ster will get different results following different 

 forecasting methods. Comparisons of the results of dif- 

 ferent methods by Rattray and Burt (1956) and by 

 Walden and Farmer (1957) will be cited here. 



(a) Rattray and Burt (1956). Rattray and Burt pre- 

 sented a hindcast of an almormally severe storm in the 

 North Pacific (.)cean. The forecasting methods of 

 Sverdrup and Munk (1947), of Bretschneider (1952) and 

 of Pierson, Neumann and James (H) were used, and the 

 results of each were compared with obser\'ations of the 

 weather ship which was in the storm area. Fig. 57 

 shows the wintl condition and the results of the compari- 

 son in the form of significant wave heights and significant 

 periods. These are defined as the average height and 

 period of one-third highest waves. 



The highest significant waves were apparently- cor- 

 rectly computed by all methods. Hindcast time histor- 

 ies of wave de\'elopment and decay de^-iate considerably 

 from the observed ones. The wave-period hindcast 

 following Sverdrup and Munk is best. 



The word "apparently" was used advisedly in the 

 first sentence of the foregoing paragraph. In formulat- 

 ing the spectra and forecasting methods the air-water 

 temperature differences are a\'eraged out and usually are 

 not considered specifically. In the application of a 

 forecasting method to an individual case, however, it is 

 advisable to consider the temperature difference since it 



