K for use in simulating all cases. To obtain best overall agreement between 

 simulated and measured profiles with constant K , the model was calibrated 

 for seven of the cases with respect to the total sum of squares. After the 

 optimal value had been determined for K , the model was verified through use 

 of two independent cases. One case (Case 700 with K = 0.3 10"^ m''/N) was 

 eliminated from the overall calibration process since water was released from 

 the tank during the run (probably to reduce wave overtopping) , lowering the 

 water level by 0.3 m, thus contaminating the case for the purpose here (see 

 Kraus and Larson 1988a). The sum of squares of the difference in depth 

 between measured and simulated beach profiles was calculated for all profile 

 surveys for all seven cases. Figure 59 illustrates the total sum of squares 

 for all cases as a function of the transport rate coefficient. A minimum 

 occurred around the coefficient value 1.6 10"^ m*/N. 



il 740 

 TO 



cr 



CO 



CO 



"to 

 o 



Calibration 



Transport Rate Coefficient (mVN)*10^ 



Figure 59. Optimization for model calibration 



180 



