On the extraction of tidal information from measurements covering a fraction of a day


  • David J. Crary,

  • Jeffrey M. Forbes


The errors incurred by extracting tidal information from data covering only a fraction of a day are investigated quantitatively as a function of fit span and tidal period. Assuming a noise component in the observational data with standard deviation σ, errors in the diurnal and semidiurnal Fourier components equal to σ occur at fit spans of 15.5 and 13.5 hours, respectively, for hourly data points. Errors increase at an exponential rate for shorter fit spans. Assuming nominal values of tidal amplitudes and standard deviations applicable to 12 hours of daytime mesopause region winds that might be obtained from MST radar measurements, it is shown that data is required at approximately 5-minute intervals to determine with acceptable accuracy tidal Fourier components from a single day's measurements. Alternatively, one to three weeks of daytime data at 30-minute intervals would enable determination of diurnal and semidiurnal components to roughly the same accuracy.