Experimental evidence for anomalous scale dependent cascading process in turbulent velocity statistics

We use a wavelet-based deconvolution method to extract some multiplicative cascading process from experimental turbulent velocity signals. We show that at the highest accessible Reynolds numbers, the experimental data do not allow us to discriminate between various phenomenological cascade models recently proposed to account for intermittency and their log-normal approximations. We further report evidence that velocity fluctuations are not scale invariant but possess more complex self-similarity properties that are likely to depend on the Reynolds number. We comment on the possible asymptotic validity of the multifractal description. (C) 1999 Academic Press.