VI. The analysis of tidal observations

The Tidal Institute was founded in the year 1919 and researches on tidal problems have since been continuously carried out. It was early shown that certain methods of analysis were incomplete and that the harmonic constituents taken into account do not fully represent the tidal oscillation. Special attention was called to the great increase required in the number of higher harmonics (“overtides” and “compound tides”) needed to represent tidal oscillations in shallow water. An investigation on the variations of “constants” obtained from yearly batches of observations provided additional evidence concerning the defects of analysis, and it was shown that the constants for important constituents are considerably perturbed by contributions from other constituents. In the year 1921 was published a very thorough expansion of the tide-generating potential, and many new constituents were indicated as being worthy of attention. Meteorological perturbations of sea-level and tides have also been studied, but much yet remains to be done, and such investigations require exact and complete methods of analysis. A tidal record may be assumed to consist of three parts:— (1) Oscillations of known periods and whose relative importance is known; (2) Oscillations whose periods are not known a priori, though exact periods may be deduced from considerations of causes, if such become known; (3) Oscillations of no persistent periodicity or amplitude and which may be regarded as sources of “casual errors" such as may be attributable to meteorological variations.