Local dimensionality and inverse persistence analysis of atmospheric turbulence in the stable boundary layer

International audienceThe dynamics across different scales in the stable atmospheric boundary layer has been investigated by means of two metrics, based on instantaneous fractal dimensions, and grounded in dynamical systems theory. The wind velocity fluctuations obtained from data collected during the CASES-99 experiment were analyzed to provide a local (in terms of scales), and an instantaneous (in terms of time) description of the fractal properties and predictability of the system. By analyzing the phase space projections of the continuous turbulent, intermittent and radiative regimes, a progressive transformation, characterized by the emergence of multiple low-dimensional clusters embedded in a high-dimensional shell and a 2-lobe mirror symmetrical structure of the inverse persistence, have been found. The phase space becomes increasingly complex and anisotropic as the turbulent fluctuations become uncorrelated. The phase space is characterized by a three-dimensional structure for the continuous turbulent samples in a range of scales compatible with the inertial sub-range, where the phase space-filling turbulent fluctuations dominate the dynamics, and is low-dimensional in the other regimes. Moreover, lower dimensional structures present a stronger persistence than the higher dimensional structures. Eventually, all samples recover a three-dimensional structure and higher persistence level at large scales, far from the inertial sub-range. The two metrics obtained in the analysis can be considered as proxies for the decorrelation time and the local anisotropy in the turbulent flow