where A(Ki), A(Oi), etc., represent the amplitude of constituents, such as Ki, Oi. The 

 type of tide is specified as 



semidiurnal, if R < 0.25, 



mixed, if 0.25 < R < 1.50 , and 



diurnal, if 1.50 < R 



The constituents identified by Ki and Oi in equation (22) are generally the dominant 

 components of the diurnal tide; constituents identified by M2 and S2 generally indicate 

 the largest components of the semidiurnal tide. 



Although the 37 terms hsted in Table 2 are adequate for most tide stations in tlie United 

 States, Zetler and Cummings (1967) have shown that 114 constituents are needed for 

 Anchorage, Alaska. A similar number of constituents are needed to describe the tides at 

 several European ports. 

 7. Tide Analyses. 



The harmonic constants of the tide, the ampUtudes A^s, and the phases k^s, of 

 equation (2) must be determined empirically from the analysis of tide records. Until the 

 early 1960's, the constants had to be derived by combining the hourly tidal heights in 

 special ways for each constituent or series of constituents as described by Schureman 

 (1941). Harmonic constants, based on 369 days of tide observations and corrected for the 

 influence of astronomical phenomena with longer periods, are generally used by tide 

 prediction agencies of other nations for practical tide predictions. The probabihty that a 

 significant error might result from the actual pattern of storms during the period of data 

 analyzed does not appear very great (it apparently has not been given much attention). It is 

 unlikely that a standard analysis would ever be based on a period which contains a major 

 storm surge. Harris, Pore, and Cummings (1965) reported that a full set of coefficients could 

 be obtained at one pass by a least squares analysis of the tide records for a period of 369 

 days in terms of the frequencies indicated by the theoretical analysis. 



At many locations, the water level variations near the coast frequently show a diurnal 

 cycle due to land-sea breezes and annual and semiannual cycles due to seasonal changes in 

 temperature and prevailing winds. It would be difficult, in the empirical analysis, to separate 

 the average value of these meteorological cycles from the gravitational cycles with the same 

 periods and there is Uttle practical reason for doing so. Both the meteorological and 

 gravitational cycles with periods of 1 day and 1 civil year (approximately an anomaUstic 

 year) may be properly called astronomical cycles as both are controlled by astronomical 

 factors. The meteorological component of the annual cycle is generally much larger than the 

 gravitational component. 



The harmonic constants for a tide station may be altered when the character of tlie 

 channel between the tide station and the open sea is changed by dredging, silting, or 



35 



