configurations, including one similar to the Kahului breakwaters. Their tests, as 

 well as others by Blue and Johnson (1949), show that the relative rotation of the 

 breakwaters makes little difference in the diffraction pattern for wave periods of 

 concern here. However, the presence of rubble on the inside of the breakwaters 

 causes significant absorption and dissipation relative to the classical vertical wall 

 breakwater. Appropriate adjustment factors taken from Bowers and Welsby 

 (1982) were applied to the spectral diffraction results of Goda (1985) for 

 comparison to two of the field calibration events. 



The calibrated short 

 wave model was run in a 

 spectral mode for the T p 

 and 6 m combinations 

 represented in the field 

 data summaries of 

 Appendix B. Comparing 

 the model results to field 

 data validates the numer- 

 ical model against an 1 1- 

 month summary of gage 

 data (Figure 42). The 

 validation comparison is 

 generally comparable to 

 the calibration results. 

 The agreement at Pier 2 

 is excellent The model 

 shows a persistent 

 tendency to under- 

 estimate the amplification 

 factor at the other gages. 

 The tendency is more 



evident than in the initial storm calibration, particularly at Canoe Club and Back 

 Basin. In most cases, it is greater than one standard deviation of the field data. 

 The possibility of incorrect reflection coefficients and/or bathymetry in these 

 shallow areas was explored within reasonable ranges, but the general level of 

 agreement could not be improved. 



Long waves. Long wave calibration was aimed at adjusting bottom friction /3 

 to approximately match amplification factors between model and data The 

 reflection coefficient K r was set to 1.0. Only the lower frequencies (0.003- 

 0.010 Hz or 100- to 333-sec period) were considered because most prominent 

 resonant peaks are in this range and K =1.0 is more strictly correct at low 

 frequencies. Only resonant peaks were considered in calibration because they 

 are the features of greatest interest and are most sensitive to the choice of f3. A 

 value of /5=0.032 was found to give a reasonably good match at all peaks in the 

 selected frequency range and at all harbor gages, as illustrated in Figure 43. 



Figure 42. Model short wave validation to 

 1 1 months of gage data 



58 



Chapter 4 Numerical Model 



