assume that the bottom contours are straight and decreasing. To estimate the 

 magnitude of the effect of refraction due to this depression, a spectral re- 

 fraction model was run for both the existing bathymetry at the FRF and a 

 smoothed bathymetry (depression removed) for deepwater wave conditions selected 

 from the set of data analyzed for this study. The results show that for the 

 Baylor gages, where the models overpredict, the correction to measured wave 

 height for this refraction would be less than 10 percent. This is much less 

 than the overprediction of the models, which is about 50 percent. For the 

 Baylor gages in the shallowest water depth, where the irregular wave model tends 

 to underpredict, the refraction correction increases the measured wave height 

 from 10 to 50 percent, which makes the disagreement between the irregular wave 

 model and the measurement even greater. 



The second factor, which may contribute to the differences between predic- 

 tion and observation, is the assumption of a Rayleigh distribution in the cal- 

 culation of significant wave height for the measured data. If the wave heights 

 do not have a Rayleigh distribution then the measured 4a will not be equal to 

 the average of the one-third highest waves. To get an estimate of how large an 

 error results from using the ho statistic for the measured H s , the individual 

 wave heights were counted for selected time series from the data set. From 

 this count, the average of the one-third higher waves was obtained. The dis- 

 tributions, which were obtained, showed that the wave heights for the most part 

 were very nearly Rayleigh distributed. The 4a values for H s differed from 

 that obtained by counting waves by less than 10 percent; for a majority of 

 cases the differences were less than 5 percent. Correction of the measured 

 wave heights for errors due to using 4a for significant wave height does not 

 significantly affect the results of the comparisons between wave measurements 

 and the model predictions. For most cases this correction tends to increase 

 the discrepancy between measurement and irregular wave model prediction. 



The comparisons showed that the irregular wave model is better than the 

 SPM-McClenan technique for predicting nearshore wave conditions. The irregular 

 wave model is more representative of the physics of the wave processes, and it 

 can be used shoreward of the SPM-defined wave breaking point, which assumes a 

 single sinusoidal wave. It also gives results which, on the whole, are in better 

 agreement with measurements than the SPM-McClenan technique. However, the com- 

 parisons of the irregular wave model results with the gage measurements show 

 large differences. In general, as the waves enter shallow water the model over- 

 predicts. This indicates that there are significant dissipative effects that 

 are not being accounted for by the model. In very shallow water, the model 

 often underpredicts in the region in or near the surf zone. A hypothesis for 

 this underprediction is that the extra measured wave energy is due to waves 

 that have broken and re-formed, a process that is not included in the irregular 

 wave model. 



VI . SUMMARY 



Comparisons of nearshore significant wave heights as estimated by an irreg- 

 ular wave procedure and by the SPM-McClenan technique with gage measurements 

 show that the irregular wave model is an improvement over the SPM-McClenan 

 technique. Yet for many cases there remain large differences between the model 

 predictions and gage measurements. Therefore, care should be taken when using 

 the model, especially in very shallow water where the model often underpredicts 

 for cases of high deepwater waves. 



17 



