439. Two cases, one from the CE data set (Case 400) and the other from 

 the CRIEPI data set (Case 6-2), were used to verify the applicability of the 

 numerical model with parameter values obtained from the calibration. Figure 

 60(a and b) shows the results of the verification runs, illustrating the 

 development in time of the beach profile together with a comparison with the 

 measured profile at the last time-step. The wave height distribution across 

 shore at the last time -step is also shown. The volume of the main breakpoint 

 bar and the amount of erosion on the foreshore are rather well predicted by 

 the numerical model. However, the crests of the bars are located somewhat too 

 far seaward, whereas the trough is not deep enough for Case 400. In general, 

 the trough is not well reproduced in the numerical model, being less pronoun- 

 ced than for the measurements, since the slope term in the transport equation 

 (Equation 33) counteracts the seaward transport of sand on the shoreward side 

 of the bar. Elimination of the slope term, however, seriously affects 

 numerical stability, resulting in a much shorter allowable time -step in 

 relation to the length step. 



Summary 



440. The developed numerical model was calibrated and verified to 

 simulate erosional (bar-type) profiles with relatively little ambiguity in 

 determining values of the required empirical coefficients. In particular, the 

 time rate of growth, volume, and location of the main breakpoint bar were well 

 reproduced. The location of the shoreline and the steep slope of the fore- 

 shore step were also well simulated. Inclusion of avalanching was needed to 

 restrict bottom slopes to within measured angles, and an explicit slope- 

 dependent contribution to the transport rate was found to greatly improve 

 stability of the model. 



441. Importantly, in all tests run with constant incident wave condi- 

 tions and water level, the calculated profile approached an equilibrium form. 

 This property is highly desirable to represent the proper time scale of 

 profile change and to use the model in an arbitrary situation without the 

 problem of numerical stability. Sensitivity analysis of the model is 

 described in the next chapter. 



181 



