mean zero-upcrossing wave period T„, with intermediate surveys every 1,000 waves. 

 They considered 5,000 waves to be a typical storm duration. Several of their significant 

 conclusions concerning damage are paraphrased as follows: 



1 . The rate of erosion of a given riprap is, as expected, 

 strongly dependent upon "the significant wave height" 



^1/3- 



2. The rate of erosion decreases with time and hence the 

 damage history curves flatten out. At the lower wave 

 heights, the curves can become nearly horizontal, giving 

 an apparently stable riprap slope. 



3. The very long preliminary tests give no certainty of the 

 riprap eroding to a totally stable equilibrium state, even 

 with low damage rates. Thus, it is not safe to assume, as 

 is often done in regular wave tests, that a slope will erode 

 to stability. All that can be said is that the erosion rate 

 may become small enough to be ignored in practice. 



4. The method of laying the riprap has a significant effect 

 upon the damage history. 



5. The wave energy spectral shape as specified by the 

 width parameter, e, does not affect the ultimate erosion 

 volume. 



6. The failure criterion requiring a given area of exposed 

 filter layer was easier to assess than that requiring the 

 observation of the erosion of filter material and gave 

 erosion volumes at failure which were independent of the 

 filter grades used. 



7. Experimental time hmited the maximum value oiN„, 

 the average number of waves incident on the riprap, to 

 5000, which is typical of a storm. In most cases this was 

 too few waves to determine whether or not equilibrium 

 damage was achieved or whether, at a given value of Hy^, 

 the slope protection would eventually fail. 



43 



