considerable evidence indicating that the MSL shoreline position effectively 

 acts as a pivot point. 



38. Figure 9 shows the distribution of median shoreline changes by 

 storm and locality.* For each storm and locality, a small "box plot" illus- 

 trates the distribution of measured profile changes. The portion of the box 

 above or below the dashed line, respectively, indicates either accretion or 

 erosion. Overall, the average median shoreline change was small, only -0.9 

 m, with a mean hinge range (difference of the hinges) of 4.8 m (+1.7 to -3.1 

 m) . Median shoreline position changes with an absolute value less than 2 m 

 were recorded for 39 percent of the cases. Only 8 percent of the shoreline 

 changes had an absolute value greater than 10 m. It is interesting to note 

 that the range of variation between cases (indicated by the heights of the 

 boxes in Figure 9) is relatively small and, with a few exceptions, measured 

 variations are as similar between storms at one locality as they are between 

 localities. Although shoreline position is traditionally used for computing 

 long-term erosion rates, based on this data set use of the change in 

 shoreline position for quantifying storm erosion appears to be limited. 

 Volume changes 



39. Figure 10 plots the distribution of volume change above MSL for 

 each locality and storm. Unlike the shoreline changes, the volume changes 

 show more consistent erosion and more variation between storms. Misquamicut 

 and Ludlam Beach had the smallest ranges in variation between profile lines; 

 whereas Nauset Beach and Jones Beach had relatively large ranges. Individual 

 profile change, represented by the extreme values, can be quite large, up to 

 -150 m^/m for one Atlantic City, New Jersey profile. Though it is difficult 

 to intercornpare storms, the data in Figure 10 illustrate the large amount of 

 variation which naturally occurs both along a beach and between beaches sub- 

 jected to the same storm. 



40. In order to account for this natural variation caused by storms, 

 the data shown in Figure 10 were used to compute a "variability factor" or 



The median value is preferred over the mean of a number of profile lines 

 because it is insensitive to single profiles with extreme changes and should 

 be more representative of the overall locality change. Plotted changes show 

 both upper and lower extreme values and "hinge" values. Hinge values are 

 defined as the 25 and 75 percentiles for the profiles at each locality. 



20 



