The current picture of wave-field development is complex. Energy from the 

 wind is transferred to intermediate and short waves in the spectrum. The 

 energy in these waves serves as a pool from which the wave-wave interactions 

 draw energy, resulting in the growth of the longer waves in the spectrum. The 

 dominant wave energy in a growing sea is seen to shift to lower frequencies. 



Often the sea is made up of a number of different wave trains. If there 

 is any significant wind, a wind sea will develop. It is initially composed of 

 short waves, but with time the wind sea waves become longer and eventually may 

 be the same length as the preexisting wave trains. If the wind is at an angle 

 different from the direction of propagation of the existing wave trains, the 

 sea surface can appear quite irregular. If the difference between wind 

 direction and the direction of propagation of the preexisting waves is small, 

 then wind seas can override the existing waves which then disappear. Often 

 the wind field is not uniform. If the wind field is curved, then the sea 

 surface can be a mixture of waves from different directions due to the same 

 wind field. Storm systems may move faster than the surface wave energy 

 generated by the storm; as a result, wave energy can be left behind by one 

 part of the storm while local generation is occurring again. Consequently, 

 wave prediction in larger waterbodies is best accomplished using numerical 

 prediction schemes. Simplified wave prediction formulas should be used only 

 in cases where the presence of energy from other wave trains can be neglected. 



2. Verification of Wave Hindcasting . 



Inoue (1967) prepared hindcasts for weather station J (located near 53° 

 N., 18° W.), for the period 15 to 28 December 1959, using a differential 

 equation embodying the Miles-Phillips-Hasselmann theory to predict wave 

 growth. A comparison of significant wave heights from shipboard observations 

 and by hindcasting at two separate locations near the weather ships is shown 

 in Figure 3-9. The calculations required meteorological data from 519 grid 

 points over the Atlantic Ocean. The agreement between observed and computed 

 values seems to justify confidence in the basic prediction model. Observed 

 meteorological data were interpolated in time and space to provide the 

 required data, thus these predictions were hindcasts. Bunting and Moskowitz 

 (1970) and Bunting (1970) have compared forecast wave heights with obser- 

 vations using the same model with comparable results. 



Wave hindcasts were developed by the U.S. Army Engineer Waterways Experi- 

 ment Station using models developed by Resio and Vincent (1977) and Resio 

 (1981). These models were based on the Miles-Hasselmann mechanisms and 

 demonstrate skill in both Great Lakes and oceanic conditions (Fig. 3-10). The 

 results of these models and the results from similarly formulated models 

 (Hasselmann et al., 1976) suggest that deepwater waves can be estimated 

 reasonably well if adequate meteorological data are available. 



3. Decay of a Wave Field . 



Wind energy can be transferred directly to the waves only when the 

 component of the surface wind in the direction of wave travel exceeds the 

 speed of wave propagation. Winds may decrease in intensity or change in 

 direction to such an extent that wave generation ceases, or the waves may 



3-21 



