Each of these areas is discussed below. 



(1) The case of the relative wavelength. In many of the laboratory 

 tests the wave period was varied to cover the range from shallow-water 

 long waves to deepwater short waves. Comparison of laboratory data 

 and MADSEN computer program predictions shows excellent correspondence 

 for shallow-water waves; e.g., at d/gT^ = 0.0065 (Table 4). As the 

 relative depth becomes larger (the wavelength becomes shorter) , the 

 computer program slightly overpredicts the observed transmission 

 coefficient (Fig. 31). This means that the prediction method is 

 conservative. Although the absolute value of the overprediction is 

 small, the percent overprediction may be large (Table 4). 



Table 4. Hffect of relative depth on prediction of Ky^. 



d 



0.0065 



0.016 



0.055 



Observed Predicted 



0.34 

 0.46 

 0.13 



0.33 

 0.44 

 0.21 



Pet. 

 error 



-3 



-4 

 ^60 



Relative 

 depth 



Shallow 



Transitional 



Docp 



B: 30cm 



I on 1.5 Slope 



0.5 



^BW12, d /h = 0.64, H/gT^ = 0.0015. 



The ability of the model to predict wave transmission coefficients 

 for a breakwater constructed entirely of armor stone is shown in Figure 

 32; wave transmission coefficients for a breakwater with a front-face 

 slope of 1 on 2. '6 are shown in Figure 33. 



(2) The case of waves breaking on the breakwater. It was difficult 

 in the laboratory to generate long waves that would break on a rough 

 permeable structure without any overtopping. However, several tests 

 that met these conditions were run using nonsurging, breaking waves 

 (Galvin, 1968). These laboratory tests show that for breaking and 

 nonbreaking waves the coefficient of transmission decreases gradually 

 as the incident steepness increases (Fig. 34); no difference was 

 evident between Ky^ for breaking and nonbreaking waves. The same 

 trend is observed in Bottin, Chatham, and Carver's (1976) data for a 

 breakwater with dolos armor units. Comparison of observed and predicted 

 coefficients of transmission through the structure shows good agreement 

 for the few breaking wave conditions tested (Fig. 34). These few tests 

 suggest that the Madsen and White (1976) model can be used for breaking 

 as well as nonbreaking waves. 



44 



