these data were used to create the synthetic northeaster events in place of 
the results from the FIMP global simulations. The maximum surge (total water 
level minus predicted tide) for each historical storm is listed in Table 12. 
196. Astronomical tides were predicted at Sandy Hook using tidal con- 
stituents provided by NOS. A 19-year tidal cycle was generated for each of 
the two storm types. Hurricane season was defined as May to October and 
northeaster season as October to April. Tides for hurricanes were predicted 
at 20-minute intervals, and tides for northeasters were predicted at 60-minute 
intervals. 
197. To develop a data base for boundary conditions to drive the proj- 
ect model, two separate sets of surge plus tide event ensembles (northeaster 
and hurricane) were created by convolving each storm time-history from the 
global FIMP model with a tide time-history. The convolution process entailed 
the superposition of each surge time-history on a predicted seasonal tide 
record at Sandy Hook, starting at the beginning of the storm season and con- 
tinuing through the season advancing the starting position by one hour. The 
maximum total water level during the surge event is recorded for each hour 
shift. To reduce the amount of data to be handled, a random selection proce- 
dure was developed to select 500 tide combinations for each northeaster 
(13,500 events) and 100 tide combinations for each hurricane (97,200 events). 
The recorded maximum water levels for each ensemble (keeping hurricane and 
northeaster events separate) were ranked by magnitude. 
198. Just as in FIMP, these two very large event ensembles could not be 
simulated by the storm surge model. Therefore, events were selected according 
to maximum water level from the ranked files described in the preceding para- 
graph. These selections were made to duplicate the separate stage-frequency 
curves that were created in FIMP for northeaster and hurricane surge plus tide 
at Sandy Hook (Figure 50). The probability of a selected event is equal to 
the proportion of the probability mass which it represents. Events were 
selected in sets of 20; two sets of hurricane and two sets of northeaster 
events. By simulating the two sets of a storm type separately and creating 
separate stage-frequency curves, a measure of the variation in the selection 
process can be determined. 
107 
