APPENDIX G 



MODEL VERIFICATION 



Prior to the surge computations the conservative properties of 

 the numerical model were examined. The model was applied first to 

 seiches in an enclosed square basin 12 kilometers on a side and 5 

 meters deep. The number of grid increments along each side was 20. 

 To invoke the most severe seiching mode, the water surface was 

 initially assigned a linear slope downward from one end of the basin 

 (j=JM) to the opposite end (j=l). The initial setup for the nonro- 

 tating basin computations was 50 centimeters, but this was reduced 

 to 10 centimeters when rotation was allowed to suppress unnecessary 

 fluctuations. The time step taken for all test runs was 50 seconds. 



The simulation of a uninodal seiche in a nonrotating basin was 

 considered first. With the specified time step, tile period of the 

 severest mode is theoretically 68.57 At. 



Figure G-l shows that the computed fundamental period is within 

 one-time increment of the theoretical period. 



90 120 

 Time (nAt) 



150 



180 



210 



Figure G-l. Computed water level at grid point (21,20) 

 for uninodal seiche test case. 



The time variation of the volume, potential, kinetic, and total 

 energies are given in Figure G-2. Each of these quantities are area- 

 weighted sums. This evaluation scheme should exhibit small perturba- 

 tions more readily than unweighted sums of the variables over the grid. 

 The variation of the volume, potential, kinetic, and total energies is 

 small, with the potential energy showing the largest deviation. 



156 



