periodic transport through the ports. 



Table 2 summarizes the peak surge at the digitized coastline for 

 each storm. Locations where the peaks occur are given in terms of 

 the grid point (I, J). The peaks and periods (Tq) of the t?q signal 

 obtained from each case are given in columns 7 , 8 and 13 , 14 . The time 

 lag of the first maximum of tjq after the center of the storms entered 

 YucateUi Strait, 6q, is also given in Table 2. 



5. Long period variation of tjq 



a) The 3.4 day volume mode 



The riQ signals obtained from all the synthetic storm simulations 

 have a long period variation superimposed on their average 28 h 

 period. This slow variation is not readily observable except for the 

 one generated by HUR5. Therefore, model HUR5 was used to investigate 

 this long period variation of the Helmholtz mode (rjg). This 

 simulation used the same storm characteristics of the original HUR5 

 hurricane. However, forcing was allowed only over the first 3 days 

 22 h of the simulation. The calculations were continued for a total 

 of 21 days to provide a clear history of the free modes in the Gulf 

 and Cayman Sea. Time histories of the mean external height cinomalies 

 of the two basins over the full simulation period are shown in Fig. 

 94. The arrow indicates when all forcing was set to zero. At this 

 time HUR5 was centered at 25.44''n and 90.86 W. 



The most striking feature of this simulation is the very long 

 mode which dominates the Cayman response. The period of this mode is 

 approximately 3.4 days. The half -cycle means of jjq over the last 7 



156l 



