After removing several outlying points from the data set, second-order 

 regression curves were determined and are also shown in Figure 13-* 

 Some of the scatter in results is probably due to Lp varying from 5.75 

 m to 15.2 m. However, there is insufficient data at overlapping values 

 of Hj. to determine a relation similar to that between F^ and 

 Lp/Df as done for Pipe-Tire floating breakwaters in Harms et al . 

 (1981). 



Results from the 6-module-beam two-dimensional prototype -scale 

 tests by Giles and Sorensen (1978) in 4 m of water are compared with the 

 La Salle Park results in Figure 14. For the Giles and Sorensen (1978) 

 data, the height H of regular waves has been substituted for H,;. 

 Agreement is surprisingly good. A second-order regression analysis of 

 the 64 field test points and 39 model test points gives 



[8] (F„.„cose)/JL = -346 + 8.76 H^ + 0.0798 (H^)^ 



MiaX C C 



where (^max^^^® ^/^ ^'^ ^'" Newtons per metre length and H^, is in 

 centimetres. The corresponding value of the square of the correlation 

 coefficient is 0.96. Equation 8 should only be used for values of 

 Hj. > 40 cm. It is \/ery similar to the relation proposed by Harms and 



* A plot of (Fmax*-°^Q^/^ versus H^, where H^ was calculated 

 according to linear wave theory (i.e., N = 1.0) showed almost no 

 difference in the regression curve over the range of data. 



94 



