Multi-Fractal Analysis of Nocturnal Boundary Layer Time Series from the BoulderAtmospheric Observatory

摘要

Time series from a nocturnal boundary layer are analyzed using fractaltechniques. The behavior of the self-affine fractal dimension, D is found to drop during a gravity wave train and rise with turbulence. D is proposed as a time series conditional sampling criterion for distinguishing waves from turbulence. Only weak correlations are found between D bulk turbulence measures such as Brunt-Vaisala frequency, Richardson number and buoyancy length. The advantages of analysis over turbulent kinetic energy (TKE), its component variances, FFT spectra, and self-similar fractals are also discussed in terms of local versus global basis functions dimensional suitability, noise, algorithm complexity, and other factors. D was found to be the only measure capable of reliably distinguishing the wave from turbulence.