Abstract

The Ondelette transformation is a mathematical tool used for time-series filtration, in an efficient numerical and computational form, by contrasting with other mathematical filters based on moving averages, Fourier transformation, etc. The main feature of this technique is that permits an eminently local analysis, isolating and improving the study of transient phenomena, which are normally showed in signal irregularities of time-series. The main objective for the use of this mathematical tool, is to remove the time-series noise, without important information loss. In this paper, the Ondelette transformation was used, related to the detection of periodic processes in meteorological time-series, e.g., surface sea breeze (SSB). The wind time-series data set (speed and direction), were obtained from the central Chile coastal zone, near Punta Hualpén (36° 45' S; 73° 11' W), located approximately 10 m on the sea surface level, and Santa Maria island (37° 02' S; 73° 30' W) at 50 m, on the sea surface level. The data set were recorded in 1991, during the summer period, when the southwesterly winds are dominant. The methodology used in this paper included a time-series filtration, using a Mallat's algorithm (orthonormally Ondelettes) and, afterwards, to carry out a basical statistic analysis, time-series rotational spectrum analysis (as a classical form), and to plot SSB's ellipses for hodographic representation. Results reveal peaks of maximum energy associated with both diurnal (±0.042 cph) and semi-diurnal (±0.084 cph) frequencies, showing SB and terrestrial breeze processes. The analysis of data showed the highest peaks associated to positive frequencies, revealing an anti-clockwise gyre of the wind, showing the influence of Coriolis force on the SSB's flux orientation.