Reprinted from MONTHLY WEATHER REVIEW.
~~~Sa.-Iy
Vol.
121.
No.
10. October
1993
An Introduction to Wavelet Analysis in Oceanographyand Meteorology: With Application to the Dispersion of Yanai Waves
S. D. MEYERS. B. G. KEU.Y, ANDJ. J. O'BRIEN
2858
MONTHLY WEATHER REVIEW
VOLUME
121
An Introduction to Wavelet Analysis in Oceanographyand Meteorology: With Application to the Dispersion of Yanai Waves
S. D. MEYERS, B. G. KEu.Y, AND J. J. O'BRIEN MesoscoleAir-Sea Interaction Group, The Florida StQle Unillersity. Tallahassee.Florida (Manuscript received30 ~ 1992,in final form 4 May 1993) ABSTRACT Wavdet analysisis a relatively newtechniquethat is an important addition to standardsignalanalysismethods. Unlike Fourier analysisthat yields an averageamplitude and phasefor each bannonic in a dataset.the wavelet transform producesan "instantaneous" estimate or local value for the amplitude and phaseof each bannonic, This allows detailed study of nonstationary spatial or time-dependent signal characteristics. The wavelet transform is discussed, examplesare given, and some methods for preprocessing data for wavelet analysis are compared. By studying the ~on ofYanai wavesin a reduced gravity equatorial model, the u~fulness of the transform is demonstrated. The group velocity is measured directly over a finite range of wavenumbers by examining the time evolution of the transform. The results agreeweD with linear theory at higher wavenumberbut the measured group velocity is reducedat lower wavenumbers,POSSIoly to interaction due with the basin boundaries.
1. Introduction a. The wavelettransform The wavelet transform has shown promise in a diversity of scientific fields, but to date it has not been much used in the oceanic and atmospheric sciences. In part, this might be due to a lack of material discussingpractical aspectsof the technique. This article therefore includes an introduction to the wavelet transform as a tool for data analysis. For brevity, we shall confine our discussionto the transform of a scalar seriesf(t). Wavelet analysis is basedon the convolution off(t) with a set of functions gab(t) derived from the translations and dilations (and rotations in higher dimensions) of a mother waveletg( t), where
g",,(t)= L
al/2
It-b (-:\
a
);
I
(1)
a(>O) and b are real. Any set of functions gab(t) constructed from ( I ) and meeting the conditions outlined below are called wavelets.The convolution off(t) with the set of wavelets is the wavelet transform (WT)
Co"esponding author address:Dr. Steven D. Meyers, Mesoscale Air-Sea Interaction Group, The Ronda StateUniversity, 8-174, 020 Love Building, Tallahassee,FL 32306-3041.
This is known as the continuous wavelet transform since a and b may be varied continuously. Translation parameter b correspondsto position or time if the data is spatial or temporal, respectively. Dilation parameter a then correspondsto scalelength or temporal period. Equation (2) expands a one-dimensional time series into the two-dimensional parameter space(b, a) and yields a local measure of the relative amplitude of activity at scale a at time b. This is in contrast to the Fourier transform that yields an average amplitude over the entire dataset. Note, we have avoided the use of the words "wavelength" or "frequency" in our description of the WT. Though waveletshave a definite scale,they need not bear any resemblanceto Fourier modes(sinesand COSines). However, a co1TeSpondence between wavelength and scale a sometimes can be achieved, as discussedin section 3. To see the limitation of standard Fowier analysis and the incentive for the development of wavelet analysis, consider the time series in Fig. la that changes frequencyhalfway through the measurement.Compare that to the signal in Fig. 1b that is generatedfrom the simultaneous presenceof both frequencies.Thesetwo very different signalsyield similar power spectra,shown in Figs. 1c,d, both being dominated by the sametwo peaks.Without prior knowledge, it would be difficult to know which signal produced which spectrum, since information on signal evolution is lost during Fourier analysis.Variations of the Fourier transform have been used (e.g., Gabor 1946) in attempts to overcome this limitation, but have met only with qualified success. The WT produces "instantaneous" coefficients and
@ 1993 American Meteorolosical
Society
()cTOBE;R
1993
MEYERS
ET AL
2859
.
0 - - .. a a 1000
1 1 :! .. 1 1 10' 100 1fT'
0.00 o.~ r.I ...
producesthe local nature of wavelet analysis,sincethe coefficients Tg{b, a) are affected only by the signal in the cone of influence (COI) about t = b. In practice, the radiusof the COI is the point I tl ='c beyond which gab{x) no longer has significant value. Usually, 'c cx:a, giving rise to conelike structures in the WT in certain cases. The COI of the endpoints is an inlportant consideration and will be discussed further in section 2. ( ii) Also, g{ t) must have zero mean. Known asthe admissibility condition, this implies the invertability of the WT. That is, the original signal can be obtained from the wavelet coefficients using f{t) = .!. II C
T ,{b, a) gab dadb
a2
where
~= f
Sine window. The in- ture hasbeenshown to be dynamically similar to Yanai formation in the end regions is now lost: (a) modulus; (b) phase. waves (mixed Rossby-gravity waves) both in the obThe negativecontours are dashed. servations (Weisberg and Horigan 1981; Tsai 1990) and in the numerical models (Cox 1980; Kindle and traditional Fourier analysis. As a practical illustration Thompson 1989; Woodberry et al. 1989). The waves of wavelet analysis we will directly measure the dis- propagatewestwardand upward with a group velocity persion ofYanai wavesin an oceanmodel. The results that is eastwardand downward ( Weisberget al. 1979; indicate general agreement with linear theory except Cox 1980). The zonal phaseand group velocities are nearthe easternboundary where the wavepropagation about 33-73 cm s-J and 16 cm S-I, respectively appearsto slow. The spectral oceanmodel usedto cre- (Weisberg et al. 1979). Similar oscillations were observedalong the equatorial front in the easterntropical ate the data is described by Kelly ( 1992). Pacific using satellite imagery (Legeckis 1977). As a result of theseobservations,it was hypothesized 3. Dispersion of Yanai waves that they were generatedby a meridional shearinstaInstability waves in the equatorial regions of the bility betweenthe westward-flowing South Equatorial world's oceanshave been under study since the mid- Current (SEC) and the eastward-flowing North Equa1970s.Interest beganfollowing the observationsmade
'\ ,~
,
, , ,,
.
~
. .
1100
aX)
.~
f
j 100 \~/(
~i
~
Tn. Dhase r, 11",-
a
800
1000
"..!
~I:f} "tifJ..
FIG. 4. Wavelet transfonn of the signal in Fig. la usingf(t) that was demeanedand detrendeci;(a) and (b) as in Fig. 3.
..
'
,
,11)1., I ,' .
, ,, 1000
1"-
FIG. 6. Wavelet transform of the sisnal in Fig. S. There is mum lessdistortion in the end regions of the data. Buft'er Ienath was 100 data points.
"0 and is disregarded.Equation (AI) can also be obtained for a real monochromatic signal with any progressive wavelet. Acknowledgments.SDM is supported by the ONR Note, it follows from (2) that any linear superpoDistinguishedEducator PostdoctoralFellowship to Dr. sition of periodic modes will result in separatelocal JamesO'Brien. The theoretical work of the Mesoscale maxima, each described as above. The WT of any Air-Sea Interaction Group is CUITently being supported function by the Physical Oceanography Branch of the Office of f(x) = ~ Aflik]x (A4) Naval Researchand the Ocean Processes Branch of NASA. Discussionson wavelet analysiswith Dr. Berj nard Barnier were useful. The color plateswere created will have modulus maxima at aj = [c + (2 + C2)1/2] using the Ferret package,provided by Dr. D. E. Har- x (2k»-I. rison and Steven Hankin. Thanks also to Mark Verschell for his on-site help with the Ferret routines. REFERENCES
APPENDIX Aristov, S. N. and P. G. Frick, 1988: Large-scale turbulence in a thin layer of nonisothermal rotating fluid. Fluid Dyn.. 23,522-528. -, and -, 1989: Large-scaleturbulence in Rayleigh-Bernard convection. Fluid Dyn.. 24, 43-48. Combes,J. M., A. Grossman,and P. T chamitchian, 1989: Wavelets: Time FrequencyMethods and PhaseSpace.Springer, 315 pp. Coulibaly, M. 1992: Analyse par ondolettes: Quelques aspectsnumeriques et applications a des signaux oceaniquessimules et a I'estimation de densitede probabilite. Ph.D. thesis.L 'universite JosephFourier-Grenoble, 209 pp. Cox. M. D., 1980:Generation and propagation of 3O-daywavesin a numerical model of the Pacific. J. Phys. Oceanogr..10,217248. Duing, W.. P. Hisard, E. Katz, J. KnaUS5, Meincke, L Miller, K. J. Moroshkin, G. Philander, A. Rybnokov, K. Voigt, and R. Weisberg, 1975:Meandersand long wavesin the equatorial Atlantic. Nature. 257,280-284. Farge,M.. 1992: Wavelet transfonns and their applications to turbulence.Ann. Rev. Fluid Mech.. 24, 395-457. Gabor, D., 1946:Theory of communications. J. 1M. Electr. Eng.. 92. 429-457. Groaman, A.. and J. Morlet, 1984: Decomposition of Hardy functions into square integrable wavelets of constant shape.SIAM J. Math Ana/.. 15,723-736. Kelly, B. G.. 1992:On the generation and dispersionofYanai waves with a spectral Cbebyshev..01, solution becomesa linear rethe lation betWeenwavelet scale and Fourier wavelength,
ao=
c + (24: C2)1/~]>..0.
(A3)
Wavelet Scales and Fourier Wavelengths The relation between wavelet scale and the more common Fourier wavelength is not necessarily straightforward.For example,some waveletsare highly irregular without any dominant periodic Components. In those casesit is probably a meaninglessexerciseto fmd a relation between the two disparate measuresof distance. However, in the caseof the Morlet wavelet, which is a periodic function enveloped by a Gaussian, it seemsmore reasonable.Using g(x)
= eicxe-x 212,
we take the transform of eikoXusing (5)
Tg(b, a) = al/2
i: dkeiNcH(k)e-(ak-C)2/2o(kkG). -
This implies
I T,(b, a)12= ali(ako)12,
2866
MONTHLY WEATHER REVIEW
VOWME 121
cilations in the wcstem Indian Ocean:Model results.J. Geophys. Res..~ 4721-4736. Legeckis, 1977:Long wavesin the eastern R., equatorialPacific~: A view ftom a geostationarysatellite.Science.197, 1179-1181. Meneveau,C., 1991:Analysis of turbulence in the orthononnal wavelet rep~ntation. J. Fluid MeclI., 232, 469-520. Meyer, Y.. 1985:Principe d'incertitude, 00ses hiIbertiennes ~ et d'operateurs. Seminaire Bourbaki. Asterisque. Scx:ieteMathematique de France, Paris, France, 662-684. Moore, D. W., and S. G. H. PhiJander,1977:Modeling of the tropical oceanic circulation. The SeQ.Goldberg, E. D.. I. N. McClave, J. J. O'Brien and J. H. Steele,Ed$. John Wiley & Sons, 319361. -, and J. P. McCreary, 1990:Excitation of intermediate-frequency equatorial wavesat a westernoceanboundary: With application to the Indian Ocean. J. Geophy.Res..9sc. 5219-5231. Monet, J., 1983:Sampling theory and wavepropagation.NATO ASI Series. 1. Issues in Acoustic Signal/Image Processingand Recognition. Chen, C. H.. Ed. Springer, 233-261. Murenzi, R., 1989: Wavelet transforms associated the N-dimento sional Euclidean group with dilations: signal in more than one dimension. Wavelets: Time Frequency Methods and Phase Space,Combes,J. M.. A. Grossmanand P. Tchamitchian. Eds. Springer, 239-246.
Philander, S. G. H., 1976: Instabilities of zonal equatorial currents. J. Geophy.Res., 81, 3725-3735. -, 1978: Instabilities of zonal equatorial currents 2. J. Geophy. Res., 83, 3679-3682. Ruskai, M B., G. Bey1kin,R. Coifman, I. Daubechies,S. MaIlat, Y. Meyer, and L Raphael, 1992: Waveletsand their Applications. Jonesand BarIcu Pub1isbers, 474 pp. Tsai, P. T. H., 1990:The 26-day oscillation in the satelliteSST measurementsin the equatorial western Indian Ocean.M.S. thesis, The Rorida State UniversitY, Department of Meteorology, ISO. Weisberg, R. H., and T. J. Weingartner, 1979: InstabilitY wavesin the equatorial Atlantic Ocean. J. Geophy. Res.. 94c, 17 98518002. Weisber& R. H., and A. M. Horigan, 1981:Low ~uency variabilitY in the equatorial Adantic. J. Phys. Oceanogr..11,913-920. Weisberg,R. H., A. M. Horigan, and C. Colin. 1979: Equatorial1y traJ)I)edRossby-gravitY propagation in the Gulf of Guinea. J. Mar. Res.. 37,67-86. Woodberry, K. E., M. E. Luther, and J. J. O'Brien. 1989:The winddriven seasonal circulation in the southerntropical Indian Ocean. J. Geophy.Res.. 94C, 17 985-18 002. Zimlin, V. D. 1981: Hierarchical model ofturbulence./zv. Almos.
Ocean. Phys..17,941-949.