The capabilities of continuous wavelet-transformation for research of temporary variability of the sea level in the north-west part of the Pacific ocean are estimate. The advantage of the wavelet analysis to research the nonstationary time series in comparison with Fourier analysis are demonstrated. The application of the wavelet-analysis (Matlab version) on example of the time series altimetry information from satellites T/P, ERS-1/2 are examined. With using the wavelet spectrograms of time series of the sea level the main scales of the fluctuations are determined. The hypotheses about the mechanisms of the external delivery of energy are proposed. Based on the estimation of slopes the areas of maxima energy in the wavelet spectrograms the direction of nonlinear energy flow of oscillations on temporary scales is determined. That the energy flow from large-scale motions to small-scale is connected to nonlinear generating of turbulence are demonstrated.