adjustments desirable to maximize coverage and regional consistency may be undesirable for 

 calculating local elevation changes (Holdahl, 1975). Adjusted level surveys should, therefore, 

 be compared only if the nature of the adjustments is fully understood. 



VU. STATISTICAL DESCRIPTION OF THE PREDICTED TIDE 



1. Introduction. 



A computer program which uses equation (2) for the prediction of hourly tidal heights 

 and the times and heights of high and low waters, was developed by Harris, Pore, and 

 Cummings (1965). Pore and Cummings (1967) pubUshed a Fortran version of the program. 

 In their program the hourly tides for 1 month are predicted first, then the time and height 

 of high and low tides are determined by refining the calculations near the extreme hourly 

 values to obtain the required accuracy in time and height. Stored tables are used for the 

 trignometric functions. A version of this program used by NOS for preparing the pubUshed 

 tide tables uses 2,048 angular steps in one quadrant of 90°. Times of high and low water are 

 calculated to the nearest minute. The choice of mean water level used by NOS in the 

 program is arbitrary. In their preparation of tide tables, MLW or MLLW is used as tlie datum 

 of tabulation and tidal heights are rounded to the nearest one-tentli of a foot. When 

 predictions are made for comparison with gage records it is convenient to use the gage zero. 

 For the theoretical study in this report the local MSL is used; thus, the mean value of the 

 predicted tide is zero. 



Calculations are made for each of the 50 NOS reference stations Usteid in Table 4. Since 

 NOS pubhshes tide predictions orUy for the reference stations, predictions for other loca- 

 tions are obtained by modifying the tide predictions for the reference stations (the proce- 

 dure is explained in the NOS Tide Tables). The information needed to compute the tides 

 by equation (2) is available for only a few nonreference stations. 



To provide insight for the probable accuracy of tlie inferred tide predictions and the 

 inferred tide statistics, this report includes calculations for a few nonreference stations 

 where the necessary harmonic constants are available. These stations are listed in Table 6. 

 The accuracy of predictions for these stations is likely to be a little better than the average 

 accuracy for all stations, because at many locations the parameters used to adjust the 

 reference station values to the specific locations have been based on a series of observations 

 of 29 days. Even shorter periods of record have been used for a few locations. Two versions 

 of the tide prediction program are used; both involve changes in the NOS version to provide 

 additional output not wanted for the pubhshed predictions, and to delete unneeded output. 



2. Tide Hydrographs. 



One version of the program is used to tabulate and plot tlie predicted tides for a 1-month 

 period for each NOS reference tide station to display characteristics of the daUy and 

 monthly cycles at each location. Plots for a few nonreference stations for which harmonic 

 constituents are available are included to provide perspective for the accuracy of the tide 



