International audienceWe consider here surf zone turbulence measurements, recorded in the Eastern English Channel using a sonic anemometer. In order to characterize the intermittent properties of their fluctuations at many time scales, we analyze the experimental time series using the Empirical Mode Decomposition (EMD) method. The series is decomposed into a sum of modes, each one narrow-banded, and we show that some modes are associated with the energy containing wave-breaking scales, and other modes are associated with small-scale intermittent fluctuations. We use the EMD approach in association with a newly developed method based on Hilbert spectral analysis, representing the probability density function in an amplitude-frequency space. We then characterize the fluctuations in a stochastic framework using a cumulant generating function for all scales, and compare the results obtained from direct and classical structure function analysis, to EMD-Hilbert spectral analysis results, showing that the former method saturates at large scales, whereas the latter method is more precise in its scale approach. These results show the strength of the new EMD-hilbert spectral analysis method for data presenting a strong forcing such as found in shallow water, wave dominated situations