Statistical Dependence of Inertial Range Properties on Large Scales in a High-Reynolds-Number Shear Flow

For turbulent velocity data measured in high-Reynolds-number atmospheric surface layer, we show that the observed dependence of the inertial range statistics on large scale fluctuations is a manifestation of the mixed averaging of regions of different local Reynolds number. This dependence can be subsumed by the Obukhov conditional averaging and the second refined similarity hypothesis of Kolmogorov. We identify a subset of velocity increments which seem to possess no large-scale dependence.