that of Kitaigorodskii, Krasitskii, and Zaslavskii (1975) in that their 

 coefficient a is expected to vary with wave conditions and not remain a 

 universal constant. 



III. FIELD EVIDENCE FOR THE FINITE -DEPTH SPECTRAL FORM 



Prior to Kitaigorodskii, Krasitskii, and Zaslavskii (1975), Kakimuma 

 (1967) and Druat, Massel, and Zeidler (1969) had noted that the shape of the 

 spectrum in shallow water deviated from Phillips' (1958) form. Kitaigorodskii, 

 Krasitskii, and Zaslavskii cited evidence supporting the f~ form, as did 

 Thornton (1977) and Gadzhiyev and Kratsitsky (1978). Ou (1980) provided 

 laboratory evidence for equation (4). A review of spectra collected at the 

 Coastal Engineering Research Center's (CERC) Field Research Facility (FRF) at 

 Duck, North Carolina, and at other gages in shallow water supports a near 

 f~^ spectral slope in depths less than 10 meters for large wave energies. 



These findings indicate a further evaluation is needed of how well the 

 equation fits observed spectra. During the Atlantic Remote Sensing Land and 

 Ocean Experiment (ARSLOE) conducted in October and November 1980 at the FRF, 

 North Carolina, wave spectra were collected in 36 meters of water about 36 

 kilometers offshore of the CERC facility (Fig. 1), using the National Ocean 

 Survey's directional buoy, XERB, with accelerometer buoys in depths of 25, 18, 

 and 17 meters of water located at distances of 12, 6, and 3 kilometers offshore 

 along a line from the facility to the XERB. In addition, data from Baylor 

 gages at seven locations in 1.5- to 9-meter depths along the FRF pier were 

 collected. On 25 October 1980 a large, low-pressure system generated waves 

 with significant heights up to 5.0 meters. Data were collected continuously 

 at the XERB during the period of high waves and spectra at all gages were 

 computed every 20 minutes. 



As a test the observed spectra, E(f), were normalized to the following 

 forms 



a (f) = f^E(f) (2Tr)-'*/g2 (10) 



P 



a^ (f) = f5E(f) (2T7)'+/g2 $(m ) (11) 



h h 



^3 (f) = f3E(f)/gj^ (12) 



Equation (10) is an estimate of the equilibrium coefficient as a function of 

 frequency if the spectra follow the deepwater form. Likewise, equation (11) 

 is an estimate of the coefficient if the spectra follow the proposed finite- 

 depth form, and equation (12) is an estimate of the coefficient if the proposed 

 shallow water (u, less than 1) holds over most of the spectrum. If any of 

 these forms fit a spectrum then the corresponding function ci(f) should be 

 constant with frequency. Therefore in a regression of f against a(f), f 

 should explain no variance; consequently, the degree of fit to the spectrum 

 by each of the three forms can be estimated by how poorly f explains variance 

 in the regression and how flat the slope with f is. The regressions were 

 performed over the region from the spectral peak to twice the spectral peak 

 and the results are tabulated in Tables 1 and 2. 



