and wave transmission through the structures. Wave reflection coefficients 

 were not measured. Wave runup on dolos was observed. 



Madsen and White (1976) developed a analytical -empirical model for the 

 prediction of wave transmission and reflection coefficients for wave trans- 

 mission through subaerial rubble-mound breakwaters. The model employs the 

 long wave assumption, so predictions using their model are expected to be 

 most reliable for shallow-water waves. Comparison of the Madsen and White 

 model with physical model tests by Keulegan (1973) and Cross and Sollitt (1976) 

 shows that the wave transmission coefficient can be predicted more reliably 

 than the reflection coefficient. 



The data from independent tests of wave transmission by overtopping con- 

 ducted in this study, together with the results of Saville (1963), Lamarre 

 (1967), Goda (1969), and Cross and Sollitt (1971), are used to develop a wave 

 transmission by overtopping equation similar to one proposed by Cross and 

 Sollitt (1971). The equation is then combined with the model of wave trans- 

 mission through permeable breakwaters of Madsen and White (1976) to form a 

 generalized model of wave transmission for breakwaters. This model is verified 

 by comparing numerical and physical model results for a wide range of conditions. 



III. LABORATORY TESTING 



1 . Laboratory Test Setup . 



Laboratory tests were performed at CERC in a wave tank 4.57 meters wide, 

 42.7 meters long, and 1.22 meters deep. A part of the tank was divided by 

 four walls to form two interior test flumes, each 61 centimeters wide; the 

 remaining tank width contained a 1 on 12 absorber beach made of crushed stone 

 with a median diameter of 2.9 centimeters (Fig. 1). This arrangement allowed 

 two experiments to be performed simultaneously, and energy reflecting off of 

 the test structures diffracts out of the test flume to minimize re-reflection 

 of waves off of the generator blade. 



The laboratory breakwaters were located between stations 5 and 10 meters 

 along the flume and parallel-wire resistance gages were used to measure wave 

 conditions in the flume. Gages placed at stations 1.40, 2.35, and 2.70 meters 

 along the test flumes were used to document incident and reflected wave condi- 

 tions. One or two gages placed landward were used to measure transmitted waves 

 (Fig- !)• 



A wave absorber consisting of a crushed gravel slope covered with a 0.6- 

 meter-thick layer of hogshair was placed at the end of the test flume to absorb 

 a majority of the transmitted wave energy. The test flume was terminated 3 

 meters before the end of the wave tank to allow water overtopping the test 

 structure to escape from the flume through the absorber gravel. This arrange- 

 ment prevented the buildup of water on the landward side of the test structure. 



2 . Methods of Generating Waves . 



Waves in this facility were generated by a programable piston-type generator 

 with a mean blade position 19 meters seaward of the entrance to the test flumes. 

 A minicomputer was used to produce monochromatic waves of a specified wave 

 height and period by moving the blade with a sinusoidal motion. Irregular waves 



13 



