(c) Survey filter layer surface. 



(d) Place riprap armor stone by dumping from a hand-held can to 

 simulate the prototype procedure of dumping from a skip. 



(e) Survey riprap armor layer surface (reference survey) . 



(f) During the generation of the predetermined number of wave 

 bursts, collect wave data and visually observe the behavior of the 

 riprap and wave runup on the riprap surface. 



(g) Survey riprap armor layer. 



(h) Increase wave height approximately 10 percent. 



(i) Repeat steps f, g, and h until failure. 



(j) Conduct final riprap survey. 



III. METHOD OF DATA ANALYSIS 



Damage to the riprap armor layer was quantified by comparing the profile of 

 the riprap armor layer taken at some wave height (damage profile) with the 

 profile taken before any waves had attacked the riprap armor (reference profile) 

 The comparison is shown schematically in Figure 4. The change in the reference 

 profile typically consisted of an erosion zone and an accretion zone, as shown 

 in Figure 4. The volume per unit length of the erosion zone was used to quan- 

 tify the extent of damage to the riprap, D. Using the median stone weight, 

 W50, to characterize the size of the riprap, the dimensionless damage, D', 

 is given by 



Wr 



where Wj- is the unit weight of the riprap stone; i.e., D' is the equivalent 

 number of median size stones removed by wave attack per median stone length. 

 The word equivalent is used because D' includes about 40 percent void spaces. 



The incident wave height was made dimensionless through the use of the 

 stability number, Ns, which was developed in Hudson's (1958) study of the 

 stability of rubble-mound breakwaters. The stability number is given by 



'50 \^/^ 



- 1 



(3) 



where H is the incident wave height, and w^^ the weight of water. Since 

 freshwater (w^ = 1000 kilograms per cubic meter) and the density of the stone 

 (wr = 2707.1 kilograms per cubic meter) were the same in both prototype and 

 small-scale tests, (Wj./w^ - 1) = 1.71 for all tests. 



Data from one small-scale test (SET-1) and one prototype test (SPL-19) are 

 used in Figure 5 to illustrate typical damage trends observed in this study. 



14 



