intensity storms, whereas a 30-sec step was required for accurate simulation 
of high-intensity events. 
216. Figures 62 through 63 display example wind field patterns and 
isovelocities for synthetic storm No. 847 for simulation hours 9 and 12. 
This storm had the following parameters: DP = 2.3 in. Hg, RM = 36.0 n.m., FS 
= 19 knots, TA = -0.5 deg, and LP = (64.2 n.m., 28.0 n.m.). Similar data were 
developed on file for each storm. 
217. Two sets of 20 selected northeasters were also simulated. Simu- 
lation time for the storms ranged from 14 to 26 hr. A constant time step of 
60 sec was used for all northeaster simulations. 
Development of Stage-Frequency Curves 
218. The method used to develop the stage-frequency curves is similar 
to the one used by Hardy and Crawford (1986). For each numerical gage, an 
array of stage intervals each with a width of 0.1 ft was created. The proba- 
bility of each storm event was added to the interval bracketing the maximum 
water level computed for that event. The exceedance, or cumulative probabil- 
ity, for each interval was calculated by adding its probability to the cumula- 
tive probability of the next higher interval whose stage is one increment 
greater. 
219. Exceedance was calculated separately for each of the two sets of 
hurricanes and northeasters producing four separate raw stage-exceedance 
curves. To smooth these raw stage-frequency curves, linear regression was 
performed on the cumulative probabilities to compute a straight line through 
the data. Most linear regression lines resulted in minimum correlation fac- 
tors of 0.98 for the hurricane sets and 0.96 for the northeaster sets. Fig- 
ure 64 contains a plot of both raw and regressed stage-frequency curves for 
the numerical gage at the west end of the I-36 bridge (gage 6). 
220. Confidence in the results is obtained by analyzing the variability 
in the stage frequencies generated by each set of hurricanes and northeasters. 
Results of this analysis for gages at the entrance to the back-bay river sys- 
tem (gage 6) and in the upper reach of the Navesink basin (gage 12 near Red 
Bank) are displayed in Figures 65 through 68. By averaging the two sets of 
regressed curves for both hurricanes and northeasters, a single, more accurate 
regressed curve for each gage is obtained. These stage-frequency curves for 
125 
