variability for that parameter. Each range was divided into iOl intervals to provide 50 inter- 

 vals above and 50 below the middle interval. The lowest computed value for each parameter 

 was taken as the lower Umit of the lowest interval for that parameter. The highest com- 

 puted value was taken as the lower limit of the highest interval. Distribution functions were 

 computed for each parameter by counting the number of times a computed value fell into 

 each interval, and were then converted into probability densities by dividing the number of 

 values in each cell by the total number of values for that parameter. The cumulative proba- 

 bility that a given parameter will exceed a particular value is obtained as the sum of the 

 probability that the parameter will have that value or a higher value. The maximum value 

 of the cumulative probability is unity (1.0000). 



Calculations were made for each of the primary tide reference stations in the United 

 States (for which daily tide predictions are published by NOS) and for the comparative 

 stations listed in Table 7. A list of these stations and the relations between the tidal datum 

 planes at each station are given in Tables 4 and 7. The locations of these stations are shown 

 in Figure 1 7. Estimates of tidal height probabilities are needed for many locations for which 

 primary tide predictions are not available. It is assumed that estimates at these stations can 

 be determined with enough accuracy by adjusting the probability density distribution 

 function derived for the reference station by tlie ratio of the tidal ranges at the two 

 locations. No better solution to this problem appears to be presently available. To facilitate 

 this estimate, the computed tidal heights are expressed as fractions of one-half the mean 

 tidal range or the diurnal range. An M in Tables 4 and 6 indicates that the mean range is 

 used for normalization; D indicates that the diurnal range is used for normaMzation. Thus, 

 absolute values for any location can be obtained by multiplying the value tabulated in 

 Appendix B by one-half of the appropriate, mean tidal range or diurnal tidal range. One -half 

 of the appropriate range has been used to produce values near 1.0 for mean tidal ranges. 

 Table 8 presents the frequency distribution functions for the extreme high and low tides of 

 each month. Similar tables for all reference and comparative stations are given in Appendix 

 B. The first line of each table gives the name of the station and the dates of the epoch (i.e., 

 the 19-year period) used in the calculations. All height values in the tables are normalized 

 by one-half of the mean tidal range or one-half of the diurnal tidal range as stated in the 

 second Une of each table. The third line identifies the variable that is the monthly high or 

 monthly low waters. 



The column on the right in Table 8 gives the class interval number, running from the 

 higher value (101) to the lowest (1). Three columns present each of the variables: the first 

 gives the lower Umit of each class interval (expressed in units of one-half the mean tidal 

 range or one-half the diurnal range as indicated on line 2); the second gives the frequency 

 with which the variable feU within the indicated class during the 19-year epoch; and the 

 third gives the cumulative frequency (i.e., the frequency with which the variable was equal 

 to or exceeded the lower limit of the class interval). Data from the first and third (variable) 

 columns are plotted in Figure 31. 



73 



