2 4 6 7 

 XMgH(,)'"Sin2»i, 



Figure 6. Real test data for example problem 2 and fitted 

 equations. Compare the correct fit through the 

 origin with the customary fit. 



CORRECT RESULTS : 



Regression coefficient 

 Correlation coefficient 

 Standard error of 3 

 Test statistic for 6 



= 0.46 

 = 0.90 

 = 0.13 

 = 3.6 



Estimated residual variance S^.^ ~ 1-^ 

 CONCLUSION : The best predictor of the Komar type is: 



V = 0.46(gHb)^/2 sin 2ab 



It would be surprising to find a clear indication of whether beach slope should 

 be included in the predictor for longshore currents by evaluating such a 

 limited data set as chosen here to encourage reader computation. Indeed a 

 comparison of Tables 3 and 4 reveals no significant differences between the 

 correlation coefficients or any other test statistics. However, significant 

 differences would be expected if a large reliable data set covering a wider 

 range of conditions were compared by the methods illustrated in this report. 



18 



