Wavelet Denoising Technique: Non white Gaussian noise reduction on oceanic microstructure data

Analysis of microstructure vertical profiles CTD (conductivity, temperature and depth) is a common method for characterizing environmental turbulent fluid dynamics. One of the objectives in analyzing high-resolution CTD profiles is to identify turbulent regions (patches) within the flow. Due to the instrumental noise of CTD measurements and the environmental characteristics, it is necessary a denoising process before the data analysis. The approaches presented in the literature for turbulent patch identification are usually unable to identify patches at low-density gradient. A new method was proposed in [1] to improve patch detection at low-density gradients. The method, pointed out the influence of wavelet mother selection and the noise characterization on the final denoising results. This article introduces a procedure to obtain the optimal wavelet filters selection to reduce noise effects. In the literature the studies based in wavelet denoising are focused on reduce the effects of white Gaussian noise. In turn, in this study, the signal representing the noise is synthetically created and modelled by both flicker and white Gaussian noise. Numerical results indicate that the selected wavelet optimize the denoising method presented in [1] and show the importance of optimizing the denoising process to identify patches in a wide range of density gradients.