On application of empirical mode decomposition for turbulence analysis in open-channel flows

Large-scale coherent structures are key elements of open-channel flow turbulence, quantification of which remains elusive. In this work, we use empirical mode decomposition (EMD) to break down a velocity time series into different modes, denoted as “intrinsic mode functions” (IMFs). Analysis of velocity auto- and co-spectra indicates that large-scale (LSMs) and very large-scale (VLSMs) fluid motions are sufficiently represented by particular groups of IMFs. A correlation between LSMs and VLSMs, identified by the EMD analysis, was found to generate 7% of the Reynolds shear stresses. However, the EMD analysis of surrogate velocity signals with randomized spectral phases demonstrated that the revealed correlation is actually an artefact of the EMD approach and should not be interpreted physically.