Acessibilidade / Reportar erro

The Fast Fourier Transform and its application to tidal oscillations

Abstracts

This paper proposes a new way of tidal spectral analysis based on the Cooley-Tukey algorithm, known as the Fast Fourier Transform. The Fast Fourier Transform analysis is used to compute both the harmonic constants of the tide and the power spectrum.The latter is obtained by means of a weighted sum. A new way is also derived to obtain the formula giving the number of the degrees of freedom,on which is based the confi dence interval corresponding to the noise spectrum.


Este trabalho propõe um novo caminho para a analise espectral da maré baseada no algoritmo de Cooley-Tukey. A análise através da "Transformação Rápida de Fourier" (Fast Fourier Transform - FFT) é empregada tanto para calcular as constantes harmônicas da maré quanto para a obtenção do espectro de energia. Este é calculado por meio de uma soma ponderada. Também é dada uma nova dedução da fórmula que exprime o número de graus de liberdade em que se baseia o intervalo de confiança correspondente ao espectro do ruído. O trabalho foi redigido em inglês a fim de facilitar o intercâmbio de informações.


The Fast Fourier Transform and its application to tidal oscillations* * This is the reprint with minor corrections and improved programs. The former reprint is obsolete.

A. dos Santos Franco; Norman J. Rock** ** Visiting Professor at the Instituto Oceanográfico da Universidade de São Paulo on a Ford Foundation and Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) grants.

Instituto Oceanográfico da Universidade de São Paulo

SYNOPSIS

This paper proposes a new way of tidal spectral analysis based on the Cooley-Tukey algorithm, known as the Fast Fourier Transform. The Fast Fourier Transform analysis is used to compute both the harmonic constants of the tide and the power spectrum.The latter is obtained by means of a weighted sum. A new way is also derived to obtain the formula giving the number of the degrees of freedom,on which is based the confi dence interval corresponding to the noise spectrum.

RESUMO

Este trabalho propõe um novo caminho para a analise espectral da maré baseada no algoritmo de Cooley-Tukey. A análise através da "Transformação Rápida de Fourier" (Fast Fourier Transform - FFT) é empregada tanto para calcular as constantes harmônicas da maré quanto para a obtenção do espectro de energia. Este é calculado por meio de uma soma ponderada. Também é dada uma nova dedução da fórmula que exprime o número de graus de liberdade em que se baseia o intervalo de confiança correspondente ao espectro do ruído. O trabalho foi redigido em inglês a fim de facilitar o intercâmbio de informações.

Full text available only in PDF format.

Texto completo disponível apenas em PDF.

ACKNOWLEDGEMENTS

Grateful appreciation is extended to the "Instituto de Física", USP, for allowing unrestricted use of the IBM 360 computer installation.

Also sincere thanks are due to the "Instituto de Comunicações Elétricas" for the free use of the IBM 1130 for high speed plotting of spectra.

A special thank you to Mr. Sylvio José Correa for his valuable collaboration in the preparation of computer programs in Appendices V, VI.

LIST OF SYMBOLS USED

a,b cosine/sine component A, B cosine/sine matrices A(Θ) auto-correlation function for time lag Θ c vector denoting phase and amplitude of oscillation CS spectral estimate at frequency s.Δf (Δf = fundamental frequency) FJ nearest Fourier frequency to Jth tidal constituent J subscript denoting tidal constituent K(Θ) cross-correlation function for time lag Θ m maximum number of lags n subscript denoting Fourier number N number of values in Fourier series p index denoting Fourier number q angular frequency (speed number) Q number of tidal constituents r phase lag reckoned from the time origin R amplitude of tidal constituent s index denoting discrete frequency of spectral estimate Ŝxy cross-spectral estimate between series x and y t time v (t) gaussian noise as a function of time x (t) time series y (t) tidal heights as a function of time z index for values of discrete weighting function α damping coefficient for weighting function γ raw Fourier spectral estimate estimate ε estimate vector denoting phase and amplitude of random oscillation Θ time lag Φ, ψ weighting functions µo r.m.s. amplitude of white noise ν degrees of freedom τ sampling interval ω angular speed difference/sum ξ, η cosine/sine of residual noise

(Received 10/15/1970)

APPENDICES

SET OF PROGRAMS FOR TIDAL ANALYSIS

by the

"Instituto Oceanográfico" Method

N.B. The programs are presented separately to provide greater flexibility to the user. In practice, however, programs in Appendices I, II, III, IV, are inter-linked and executed sequentially by the computer.

Click to enlarge

APPENDIX I

FAST FOURIER TRANSFORM

Click to enlarge

APPENDIX II

MATRIX GENERATION

Click to enlarge

APPENDIX III

CALCULATION OF TIDAL COMPONENTES aj AND bj., HARMONIC CONSTANTS H&g AND CORRECTION OF FOURIER COEFFICIENTS FOR TIDAL EFFECTS TO OBTAIN RESIDUALS ξn AND ηn

Click to enlarge

APPENDIX IV

POWER SPECTRAL ANALYSIS

Click to enlarge

  • CATTON, D.B & CARTWRIGHT, D.E. 1963. On the Fourier analysis of tidal observations, Int hydrogr. Rev., vol. 40, no. 1, p. 113-125.
  • COOLEY, J W. & TUKEY, J.W. 1965. An algorithm for the machine calculation of complex Fourier series. Maths Comput., vol. 19, p. 297-301.
  • DOODSON, A. T. 1928. The analysis of tidal observations Phil. Trans. R. Soc Ser. A, vol. 227, p. 223-279.
  • FRANCO, A.S. 1966. Tides - Fundamentals. Prediction analysis. Monaco, Int. hydrogr. Bureau, 369 p.
  • ______ 1967. La méthode Munk-Cartwright pour la prediction de la marée vue de la lumière de l'idée fondamentale de Laplace C.r.hebd. Séanc. Acad. Sci., Paris, octobre 1967.
  • ______ 1968 The Munk-Cartwright method for tidal prediction and analysis. Int. hydrogr. Rev., vol. 45, no. 1, p.155-165.
  • ______ 1970 Fundamentals of power spectrum analysis as applied to dis crete observations. Int. hydrogr. Rev., vol. 47, no. 1, p 91-112.
  • GODIN, G. 1970. La résolution des ondes composantes de la marée. Int. hydrogr. Rev., vol. 47, no. 2, p. 139-150.
  • HORN, W. 1960, Some recent approaches on tidal predictions Int. hydrogr. Rev., vol. 37, no. 2, p. 65-84.
  • LENNON, G, W. 1969, An intensive analysis of tidal data in the Thames Estuary. Proc. of the Symposium on tides, Monaco, 28-29/4/1967, Unesco.
  • MIYASAKI, M. 1958. A method for the harmonic analysis of tides. Oceanogrl Mag., vol. 10, no. 1.
  • MÜNK, W.H. & CARTWRIGHT, D.E. 1966. Tidal spectroscopy and prediction. Phil. Trans. R. Soc., Ser., vol. 259, no. 1105, p. 533-581.
  • MÜNK, W.H. HASSELMANN, K. 1964. Super resolution on tides. Studies on Oceanography. In Yoshida K. ed. - Studies on oceanography. Tokyo, Univ. Press, p. 339-344.
  • MUNK, W.H., SNODGRASS, F.E. & TUCKLER, M.J. 1959. Spectra of low frequency ocean waves. Bull. Scripps Instn Oceanogr. tech. Ser., p. 283-292.
  • MUNK, W.H., ZETLER, B.D. & GROVES, G.W. 1965. Tidal cusps. J.geophys. Res., vol. 10, p. 211-219.
  • ZETLER, B.D. 1969. Shallow water tide predictions. Proc. of the Symposium on tides, Monaco, 28-29/4/1967, Unesco.
  • ZETLER, B.D. & CUMMINGS, R.A. 1967. A harmonic method for predicting shallow-water tides. J. mar. Res., vol 25, no. 1, p. 103-114.
  • *
    This is the reprint with minor corrections and improved programs. The former reprint is obsolete.
  • **
    Visiting Professor at the Instituto Oceanográfico da Universidade de São Paulo on a Ford Foundation and Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) grants.
  • Publication Dates

    • Publication in this collection
      12 June 2012
    • Date of issue
      1971

    History

    • Received
      15 Oct 1970
    Instituto Oceanográfico da Universidade de São Paulo Praça do Oceanográfico, 191, 05508-120 São Paulo SP Brasil, Tel.: (55 11) 3091 6513, Fax: (55 11) 3032 3092 - São Paulo - SP - Brazil
    E-mail: amspires@usp.br