to two factors: datum discrepancies and local superelevation. The datum for 

 all tide gages was based on the National Geodetic Vertical Datum (Nt;VD), which 

 is not intended to equal the local mean water leval. Bench marks used tc 

 reference the gaqes were not tied together by a single survey and hence ...ay 

 not lie in the same datum plane. In addition, .local mean water levels, par- 

 ticularly in the sounds, should be above NGVD due to a general rise in sea 

 level, since NGVD was established in 1929, and also due to local supereleva- 

 t.on of the bays caused by freshwater inflow into the sounds. Freshwater 

 inflow, and its local effects, was not considered in this modeling effort. In 

 order to account for the above factors, datum adjustments were estimated as 

 differences between the NGVD values and the monthly high/low averages. Table 3 

 summarizes the sites used in the verification of the two largest models, gages 

 wjtliin trie grids, and tl;e datum adjustments estimated from the high/low 

 averages. Tabulated data were not available for all of the prototype mari- 

 grams, and it must be noted that the March 1962 surge at Nags Head is esti- 

 mated since the tide gage there was lost during the storm. Grygiel states 

 that the estimate "is based on reliable evidence and should be less than 1 ft 

 in error. . ." (p. 3). 



Offshore Model 



45. Plates 11-13 compare the computed storm surge cf the March 1962 

 storm with the prototype data for the il sites listed in Table 3. Where datum 

 adjustments are available, they have been subtracted from the prototype 

 records in these plates. While the comparisons agree well, some differences 

 occur due to the shoreline approximations used in the offshore model .?,nd due 

 to the presence of wave setup in the prototype data. Computations for 

 Oriental, Rodanthe, Point Harbor, and Columbia exhibit excellent fits to the 

 field records. Minnesiott Beach and Cherry Point have computed peak sui ges 

 that are slightly high due the shorellno approximations made for the Neuse 

 River. The cell sizes in the offshore grid ner.r Stumpy Point Bay also do not 

 allow this small bay to be well represented; therefore the computations and 

 prototype for this site differ. The presence of wind-induced wave setup in 

 the prototype records of V/rightsville Beach, Hatteras, and Englehard can 

 explain why these computations fall below field measurements. The divergence 

 in elevations after hour 60 (the peak of the storm) at Hatteras shows how wave 



90 



